Y034013 1 1626066 B86014297, C86011611 Super chips for artificial intelligence Reddy, R. Carnegie-Mellon Univ., Pittsburgh, PA, USA Sponsor: IEEE; Univ. Pennsylvania 1985 IEEE International Solid-State Circuits Conference, 32nd ISSCC. Digest of Technical Papers (Cat. No. 85CH2122-0) 54-5 1985 13-15 Feb. 1985 New York, NY, USA Publ: Lewis Winner, Coral Gables, FL, USA 394 pp. U. S. Copyright Clearance Center Code: 0193-6530/85/0000-0054$01.00 Treatment: GENERAL, REVIEW Doc. Type: CONFERENCE PAPER Lang.: ENG. (2 Refs) Summary form only given. Following a discussion of the need for super chips and of the computational complexity of artificial intelligence (AI) problems, several applications for AI are given. These include applications in vision and speech, robotic manipulation and mobility, combinatorial search in game playing, expert systems, and massively parallel processing. Desc.: artificial intelligence; microprocessor chips; parallel processing; VLSI Ident.: VLSI; ULSI; microprocessor chips; artificial intelligence; super chips; computational complexity; AI; applications; vision; speech; robotic manipulation; mobility; combinatorial search in game playing; expert systems; massively parallel processing Class Codes: B1265F; B2570; C1230; C5130; C5220 Y034013 2 1625924 B86013940, C86015404 Application of artificial intelligence techniques to design for testability verification Cabodi, G.; Camurati, P.; Prinetto, P. Dipartimento di Autom. e Inf., Politecnico di Torino, Italy Luque, E. (Editors) Mini and Microcomputers and their Applications. Proceedings of the ISMM International Symposium 7-10 1985 25-28 June 1985 Sant Feliu de Guixols, Spain Publ: Int. Soc. Mini & Microcomput., Barcelona, Spain 518 pp. Treatment: THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (7 Refs) The paper presents a possible application of artificial intelligence techniques to the verification of a set of high level design rules. The rules under consideration are in the field of design for testability and concern level sensitive scan design. The artificial intelligence techniques are intelligent searching strategies. The knowledge representation which supports such an approach is presented in the companion paper 'Interpretation of a CAD system data base as a Semantic Network' (ibid., p.11). Desc.: artificial intelligence; circuit CAD; integrated circuit testing; logic testing; VLSI Ident.: artificial intelligence techniques; testability verification ; high level design rules; design for testability; level sensitive scan design; intelligent searching strategies Class Codes: B1130B; B2570; C7410D; C1230; C5210B Y034013 3 1617046 D86000495 Computer software-protective legislation at last McFarlane, G. Account. Mag. (GB) vol.89, no.954 533 Dec. 1985 CODEN: ACMAEC ISSN: 0001-4761 Treatment: GENERAL, REVIEW Doc. Type: JRNL PAP. Lang.: ENG. Nowadays software packages are designed at immense cost, and they represent a considerable investment in time and money. But they are very easily copied and clearly the copier, who has not made the same investment, enjoys a commercial advantage over the originator of the program which many would consider unfair. A Private Member's Bill was introduced in Parliament to rectify the position in the UK, and as the Copyright (Computer Software) Amendment Act 1985 it received Royal Assent on 16 July. The Act applies to computer programs the protection afforded to literary works in 1956. It restricts, inter alia, reproduction in any material form, publication, transmission and adaptation. The Act also extends to an infringing copy of a program embodied on a medium such as tape, disc or chip, wider provisions relating to criminal offences and search warrants. Desc.: legislation; security of data; software packages Ident.: legislation; software packages; Copyright (Computer Software) Amendment Act 1985; program Class Codes: D1020; D2000 Y034013 4 1614885 C86014345 A meta-level extension of Prolog Bowen, K.A.; Weinberg, T. Dept. of Comput. & Inf. Sci., Syracuse Univ., NY, USA Sponsor: IEEE 1985 Symposium on Logic Programming (Cat. No.85CH2205-3) 48-53 1985 15-18 July 1985 Boston, MA, USA Publ: IEEE Comput. Soc. Press, Washington, DC, USA x+271 pp. ISBN 0 8186 0636 3 U. S. Copyright Clearance Center Code: CH2205-3/85/0000-0048$01.00 Treatment: PRACTICAL; THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (5 Refs) A metalevel extension of Prolog is described which continues the work of K.A. Bowen & R.A. Kowalski (1982). It reifies not only theories and formulas but also proofs and search spaces (and, implicitly, substitutions), and introduces concurrency constructs. Its use in programming a metalevel approach to the problem of faulty circuit diagnosis is described. This approach appears to provide a logically sound programming formalism sufficiently powerful to write clear reliable programs for experimental and applied artificial intelligence. Desc.: artificial intelligence; logic programming; PROLOG Ident.: meta-level extension; Prolog; proofs; search spaces; concurrency constructs; faulty circuit diagnosis; artificial intelligence Class Codes: C6140D; C6110; C1230 Y034013 5 1614273 C86014710 Supportive interaction between rule based and self learning expert systems in biomedical research van Braam, W.; Dassen, W.; Brugada, P.; Gorgels, A.; Wellens, H. Dept. of Cardiology, Limburg Univ., Maastricht, Netherlands Sponsor: SCS Proceedings of the 1985 Summer Computer Simulation Conference 465-9 1985 22-24 July 1985 Chicago, IL, USA Publ: SCS, San Diego, CA, USA xxxiv+758 pp. Treatment: APPLIC; PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (1 Refs) Recently an expert system based on rules describing causal relations has been developed to facilitate the design and evaluation of cardiac pacemakers. In this system, the rules that define the behaviour of the pacemaker are separated from the rest of the program, and stored in the knowledge base. The system was also applied in other studies in which causal relations could be defined. A second expert system has been developed. It is to be used in fields where the rules describing causal relations are not well-defined and also to evaluate the relative importance of parameters that are shared by different diagnoses. To achieve these aims, the system has been made to have self-learning properties. By entering a list of parameters and a diagnosis, the weight of the different parameters that are used to determine the correct diagnosis can be assessed. By applying the same set of parameters and diagnoses to both systems, the self-learning system is able to produce correct diagnoses without the need to explicitly formulate rules. It is also possible to study the difference between the theoretical outcome (the rule based system) and the actual outcome (the self-learning expert system). Recently the rules to differentiate between different supraventricular and ventricular tachycardias in combination with a set of actual patient data were evaluated. Based on these findings an optimisation of the rules used in the first expert system could be achieved. Desc.: expert systems; medical diagnostic computing Ident.: supportive interaction; rule based expert systems; supraventricular tachycardias; self learning expert systems; biomedical research; causal relations; cardiac pacemakers; knowledge base; diagnoses; ventricular tachycardias; patient data; optimisation Class Codes: C7140; C7330 Y034013 6 1614038 C86014515 Expert systems: an overview Verdejo, F. Facultad de Informatica, Pais Vasco Univ., Bilbao, Spain Verkerk, C. (Editors) Sponsor: CERN 1984 CERN School of Computing (CERN 85-09) 161-9 1985 9-22 Sept. 1984 Aiguablava, Catalonia, Spain Publ: CERN, Geneva, Switzerland vii+376 pp. Treatment: GENERAL, REVIEW Doc. Type: CONFERENCE PAPER Lang.: ENG. (7 Refs) The purpose of this article is to introduce readers to the basic principles of rule-based expert systems. Four topics are discussed in subsequent sections: (1) definition; (2) structure of an expert system; (3) state of the art and (4) impact and future research. Desc.: expert systems; knowledge engineering Ident.: inference mechanisms; knowledge base; software tools; rule-based expert systems; state of the art Class Codes: C7000; C1230 Y034013 7 1613642 C86011613 Induction, of and by probability Rendell, L. Issued by: Univ. Illinois at Urbana-Champaign, USA; July 1985 6 pp. Report No.: UIUCDCS-R-85-1209 Treatment: THEORETICAL Doc. Type: REPORT Lang.: ENG. (25 Refs) Examines some methods and ideas underlying the author's successful probabilistic learning systems (PLS). These systems have proven uniquely effective and efficient for generalization learning (induction) in heuristic search. Aspects of PLS include use of probabilities to guide both task performance and learning, incremental revision and normalization of probabilities, and localization and correction of their errors. Construction of new terms (features) for heuristic functions may be feasible. Desc.: learning systems; probability Ident.: probabilistic learning systems; generalization learning; induction; heuristic search; incremental revision; normalization of probabilities Class Codes: C1240; C1140Z; C1230 Y034013 8 1613124 C86014307 Adding meta rules to OPS5. A proposed extension Rosenthal, D. Space Telescope Sci. Inst., Baltimore, MD, USA SIGPLAN Not. (USA) vol.20, no.10 79-86 Oct. 1985 CODEN: SINODQ ISSN: 0362-1340 Doc. Type: JRNL PAP. Lang.: ENG. OPS5 is a rule-based nonprocedural programming language rapidly gaining popularity. It is a pure 'object level' programming language, as it provides the programmer with no control constructs. The approaches taken when using OPS5 is to generate a rule base such that, during program execution the most appropriate rule will 'select itself' and be applied at the appropriate time. Because production memory (the rule base) in OPS5 is strictly segregated from working memory (data) and the conflict set (currently applicable rules), an OPS5 program cannot direct the conflict resolution strategy of the interpreter. This can lead to situations where the OPS5 programmer may feel forced to write rules which 'trick' the interpreter into boosting a rule's priority. Such artificial constructs are clearly harmful. An extension to OPS5 is proposed which allows the user to affect conflict resolution through the mechanism of meta-rules. A further refinement proposes implementing meta-rules via a programmable conflict resolution strategy. These extensions represent a considerable enhancement of the expressiveness of the present language, with little modification to existing implementations. They are in keeping with the OPS5 programming s style, and, in addition, would preserve compatibility with existing OPS5 code. Desc.: high level languages Ident.: object level language; meta rules; OPS5; rule-based nonprocedural programming language; control constructs; program execution; production memory; working memory; conflict set; currently applicable rules ; conflict resolution strategy; interpreter; expressiveness; compatibility Class Codes: C6140D Y034013 9 1612455 C86011608 Representing knowledge Gallaire, H. European Comput. Ind. Res. Centre, Munich, Germany Recherche (France) vol.16, no.170 1240-2, 1245-8 Oct. 1985 CODEN: RCCHBV ISSN: 0029-5671 Treatment: GENERAL, REVIEW Doc. Type: JRNL PAP. Lang.: French The need is explained for expert systems, designed for 'intelligent' tasks, to reproduce in part the intellectual behaviour of a human specialist handling a large mass of knowledge accumulated over a long period. A condition for success will be the discovery of formalisms which can represent the wide diversity of forms of knowledge used by human experts. A goal of high priority is the preservation of neutrality in the representation, irrespective of the eventual use to which it will be put by the inference engine. Desc.: expert systems; knowledge engineering Ident.: knowledge representation; expert systems; intellectual behaviour; human specialist; formalisms; inference engine Class Codes: C1230; C0200 Y034013 10 1608407 C86011598 Automatic theorem proving with configurations Bottger, H. Elektron. Informationsverarb. Kybern. (Germany) vol.21, no.10-11 523-46 1985 CODEN: EIVKAX ISSN: 0013-5712 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: German (39 Refs) A new non-resolution method of theorem proving is presented. The relevant information is encoded in labeled trees-so-called 'configurations'. The semantics of configurations is defined by a correspondence to first-order expressions. Inference rules on configurations and procedures working with them are introduced and proved to be correct and complete. Methods for restricting the search space are investigated. Desc.: theorem proving; trees (mathematics) Ident.: nonresolution method; inference rules; PLANNER; PROLOG; configurations; theorem proving; labeled trees; semantics; first-order expressions; search space Class Codes: C1230; C1160 Y034013 11 1607862 C86011595 Expert systems overview: where does the phrase fit? Tsotsos, J.K. Fac. of Comput. Sci., Toronto Univ., Ont., Canada CIPS Rev. (Canada) vol.9, no.5 12-14 Sept.-Oct. 1985 CODEN: CIREDA ISSN: 0701-8681 Treatment: PRACTICAL Doc. Type: JRNL PAP. Lang.: ENG. (2 Refs) Accepted definitions of expert systems tend to lack substance. The author points out two major classes of criteria which he feels are important in determining whether a computer program is an expert system. These criteria are the external characteristics of the system, namely utilities, performance and transparency to users, and the technology used to build the system, namely, appropriateness of representation, appropriateness of reasoning and control, and facilities for explanation, knowledge and performance validation. Active research concensus are listed. Desc.: expert systems Ident.: knowledge facilities; user transparency; explanation facilities; expert systems; computer program; external characteristics; utilities; performance; representation; reasoning; control; performance validation Class Codes: C1230; C7000 Y034013 12 1606771 C86013215 Processing of semantic nets on dataflow architectures Bic, L. Dept. of Inf. & Comput. Sci., California Univ., Irvine, CA, USA Artif. Intell. (Netherlands) vol.27, no.2 219-27 Nov. 1985 CODEN: AINTBB ISSN: 0004-3702 U. S. Copyright Clearance Center Code: 0004-3702/85/$3.30 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (6 Refs) Extracting knowledge from a semantic network may be viewed as a process of finding given patterns in the network. On a von Neumann computer architecture the semantic net is a passive data structure stored in memory and manipulated by a program. This paper demonstrates that by adopting a data-driven model of computation the necessary pattern-matching process may be carried out on a highly-parallel dataflow architecture. The model is based on the idea of representing the semantic network as a dataflow graph in which each node is an active element capable of accepting, processing, and emitting data tokens traveling asynchronously along the network arcs. These tokens are used to perform a parallel search for the given patterns. Since no centralized control is required to guide and supervise the token flow, the model is capable of exploiting a computer architecture consisting of large numbers of independent processing elements. Desc.: artificial intelligence; computer architecture; parallel processing Ident.: artificial intelligence; semantic nets; dataflow architectures; von Neumann computer architecture; passive data structure; data-driven model; pattern-matching process; dataflow graph; parallel search Class Codes: C5220; C1230 Y034013 13 1605813 B86011445, C86008018 Diagnosis: using automatic test equipment and artificial intelligence expert systems Ramsey, J.E., Jr. Dept. of Electr. Eng., Air Force Inst. of Technol., Wright-Patterson Air Force Base, OH, USA Sponsor: IEEE; Aerosp. Electron. Syst. Soc Proceedings of the IEEE 1985 National Aerospace and Electronics Conference NAECON 1985 (Cat. No. 85CH2189-9) 1369-74 vol.2 1985 20-24 May 1985 Dayton, OH, USA Publ: IEEE, New York, USA 2 vol. 1595 pp. Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (2 Refs) The author reviews three expert systems (ATEOPS, ATEFEXPERS, and ATEFATLAS) created to direct automatic test equipment (ATE). Each system uses a different knowledge base or inference engine, and each bases its testing on the circuit schematic, test requirements document (TRD), or ATLAS code. Implementing generalized modules allows the expert systems to be used for any different unit under test. These expert systems give some insight into the necessary knowledge bases and inference engines needed to have an expert system direct ATE. Using converted ATLAS to LISP code allows the expert system to direct any ATE using ATLAS. The constraint propagated frame system allows for the expansion of control by creating the ATLAS code, checking the code for good software engineering techniques, directing the automatic test equipment, and changing the test sequences as needed (planning). Desc.: artificial intelligence; automatic test equipment; expert systems Ident.: automatic test equipment; artificial intelligence expert systems; ATEOPS; ATEFEXPERS; ATEFATLAS; knowledge base; inference engine; circuit schematic; test requirements document; ATLAS code; generalized modules; LISP code; software engineering Class Codes: B7210B; C3380; C7410D Y034013 14 1596211 C86010467 Engineering and scientific software control standards modelling: an interactive expert system using telecommunications Rushinek, A.; Rushinek, S. Miami Univ., Coral Gables, FL, USA Adey, R.A. (Editors) Engineering Software IV. Proceedings of the 4th International Conference 4/3-16 1985 June 1985 London, England Publ: Comput. Mech. Centre, Southampton, England 1098 pp. ISBN 0 905451 33 3 Treatment: GENERAL, REVIEW; THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (43 Refs) Selecting an Engineering and Scientific software system (ESS) is an arduous process. The overall satisfaction derived from this type of system depends on many variables. This paper analyzes the influence of a number of ESS predictor variables on overall satisfaction. It confirms the theories that suggest that ESS ease of operation, computer reliability and tech support-trouble shooting are the major determinants of overall computer user satisfaction. Desc.: engineering computing; natural sciences computing; software packages; software selection Ident.: software control standards modelling; interactive expert system; telecommunications; Engineering and Scientific software system; ESS predictor variables; ESS; ease of operation; computer reliability; tech support-trouble shooting; computer user satisfaction Class Codes: C7400; C7300; C0310 Y034013 15 1595374 C86010458 The effects of technical support on user satisfaction using an interactive artificial intelligence expert system Rushinek, A.; Rushinek, S.F. Miami Univ., Coral Gables, FL, USA Sponsor: IEEE; American Autom. Control Council Proceedings of the 1985 American Control Conference (Cat. No. 85CH2119-6) 904-5 vol.2 1985 19-21 June 1985 Boston, MA, USA Publ: IEEE, New York, USA 3 vol. 1720 pp. Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (5 Refs) The authors analyze the influence of technical support (TS) on overall computer user satisfaction as determined by multiple regression. Trouble shooting, user expectations, and documentation are found to be the significant variables affecting overall satisfaction, whereas the number of computer systems and type of the microcomputer have the least significant effect in the user satisfaction model. An interactive expert system for TS is demonstrated. The implication for improvements in TS by updates to an AI database are discussed and analyzed. Desc.: expert systems; system documentation Ident.: technical support; user satisfaction; interactive artificial intelligence; expert system; computer user; documentation; microcomputer; database Class Codes: C7400; C6100 Y034013 16 1594984 C86006757 Substantial constructive induction using layered information compression: tractable feature formation in search (revised) Rendell, L. Issued by: Univ. Illinois at Urbana-Champaign, USA; May 1985 9 pp. Report No.: UIUCDCS-R-85-1198 Treatment: THEORETICAL; EXPERIMENTAL Doc. Type: REPORT Lang.: ENG. (32 Refs) Addresses a problem of induction (generalization learning) which is more difficult than any comparable work in AI. The subject of the present research is a hard problem of new terms, a task of realistic constructive induction. While the approach is quite general, the system is analyzed and tested in an environment of heuristic search where noise management and incremental learning are necessary. Here constructive induction becomes feature formation from data represented in elementary form. A high-level attribute or feature such as 'piece advantage' in checkers is much more abstract than an elementary descriptor or primitive such as contents of a checkerboard square. Features have often been used in evaluation functions; primitives are usually too detailed for this. To create abstract features from primitives (i.e. to restructure data descriptions), a new form of clustering is used which involves layering of knowledge and invariance of utility relationships related to data primitives and task goals. Desc.: artificial intelligence; learning systems; pattern recognition Ident.: hypothesis forming; inference; inductive reasoning; layered information compression; knowledge layering; generalization learning; AI; realistic constructive induction; heuristic search; noise management; incremental learning; feature formation; high-level attribute; piece advantage; checkers; abstract features Class Codes: C1230; C1250 Y034013 17 1594721 C86010451 A plant failure diagnosis expert system written in relational database manipulation languages Kumamoto, H.; Suzuki, T.; Inoue, K.; Ikenishi, K. Fac. of Eng., Kyoto Univ., Japan Trans. Soc. Instrum. & Control Eng. (Japan) vol.21, no.8 842-8 Aug. 1985 CODEN: TSICA9 ISSN: 0453-4654 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: Japanese (18 Refs) Most expert systems (ES) are programmed in artificial intelligence languages such as LISP and PROLOG. The paper proposes a new approach which programs ES for plant failure diagnosis in a data manipulation language (DML) of a relational database. The personal computer DML of dBASE II or K-MAN is used. The approach is justified, then a knowledge acquisition and representation scheme of incremental refinement type is introduced for the failure diagnosis, and the inference engine and other parts of the system are implemented accordingly by the DML. The resulting ES is overviewed, and it is demonstrated by application to two subsystems of a ship engine propulsion system. Desc.: artificial intelligence; expert systems; failure analysis; high level languages; marine systems; propulsion; relational databases; system recovery Ident.: knowledge acquisition scheme; knowledge representation scheme; plant failure diagnosis expert system; relational database manipulation languages; artificial intelligence languages; personal computer; dBASE II; K-MAN; incremental refinement; inference engine; ship engine propulsion system Class Codes: C7400; C6160D; C1230 Y034013 18 1591569 C86006742 Experiments with semantic paramodulation McCune, W.; Henschen, L. Div. of Math. & Comput. Sci., Argonne Nat. Lab., IL, USA J. Autom. Reasoning (Netherlands) vol.1, no.3 231-61 1985 CODEN: JAREEW ISSN: 0168-7433 U. S. Copyright Clearance Center Code: 0168-7433/85$0.15 Treatment: THEORETICAL; EXPERIMENTAL Doc. Type: JRNL PAP. Lang.: ENG. (25 Refs) A semantic restriction is presented for use with paramodulation in automated theorem-proving programs. The method applies to first-order formulas with equality whose clause forms are Horn sets. In this method, an interpretation is given to the program along with the set of clauses to be refuted. The semantic restriction was incorporated into an existing automated theorem-proving program, and a number of experiments were conducted on theorems in abstract algebra. Semantic paramodulation is an extension of set-of-support paramodulation. A comparison of the two strategies was used as the basis for experimentation. The semantic searches performed markedly better than the corresponding set-of-support searches in many of the comparisons. The semantic restriction seems to make the program less sensitive to the choice of rewrite rules and supported clauses. Desc.: theorem proving Ident.: semantic inference rules; semantic paramodulation; automated theorem-proving; Horn sets; interpretation; semantic restriction; abstract algebra; semantic searches; set-of-support Class Codes: C1230 Y034013 19 1591543 D86000348 The quest for the intelligent database Basu, D. ICP Bus. Software Rev. (Eur. Ed.) (GB) vol.1, no.5 12-13 Dec. 1985 CODEN: IBSEEX Treatment: GENERAL, REVIEW Doc. Type: JRNL PAP. Lang.: ENG. While there is theoretical promise in making artificial intelligence (AI) a resident component of-or normal front end for-database management systems (DBMSs), there is a lot of work ahead. Current AI technology cannot explode in a standalone environment. The merging of AI as an extension of today's database technology will make AI technology more creditable. The areas where AI can be incorporated are: ad hoc database searching; natural language facilities; graphics interfaces; concurrent query support; rule-based query support; natural language use of ad hoc query/report writer; and inclusion of Prolog in relational DBMS. It must be remembered that AI is not a panacea and, like other discrete technologies, never will be. The implementation of AI in DBMS will not solve all problems and the information industry will not be ready to accept it commercially without bona fide technical merit. Desc.: database management systems; expert systems Ident.: intelligent database; artificial intelligence; front end; database management systems; natural language facilities; graphics interfaces; concurrent query support; rule-based query support; ad hoc query/report writer; Prolog; relational DBMS; DBMS Class Codes: D2080; D2000 Y034013 20 1591431 C86009255 R1-Soar: an experiment in knowledge-intensive programming in a problem-solving architecture Rosenbloom, P.S.; Laird, J.E.; McDermott, J.; Newell, A.; Orciuch, E. Dept. of Comput. Sci., Stanford Univ., CA, USA; IEEE Trans. Pattern Anal. & Mach. Intell. (USA) vol.PAMI-7, no.5 561-9 Sept. 1985 CODEN: ITPIDJ ISSN: 0162-8828 U. S. Copyright Clearance Center Code: 0162-8828/85/0900-0561$01.00 Treatment: APPLIC; THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (19 Refs) The authors present an experiment in knowledge-intensive programming within a general problem-solving production-system architecture called Soar. In Soar, knowledge is encoded within a set of problem spaces, which yields a system capable of reasoning from first principles. Expertise consists of additional rules that guide complex problem-space searches and substitute for expensive problem-space operators. The resulting system uses both knowledge and search when relevant. Expertise knowledge is acquired either by having it programmed, or by a chunking mechanism that automatically learns new rules reflecting the results implicit in the knowledge of the problem spaces. The approach is demonstrated on the computer-system configuration task, the task performed by the expert system R1. Desc.: automatic programming; expert systems Ident.: knowledge-intensive programming; problem-solving architecture; reasoning; complex problem-space searches; chunking; expert system Class Codes: C6110; C7430; C1230 Y034013 21 1589707 C86006746 A new technique in heuristic search-algorithm SA Zhang Bo; Zhang Ling Dept. of Comput. Eng. & Sci., Tsinghua Univ., China J. Tsinghua Univ. (China) vol.25, no.2 1-14 1985 CODEN: CHHPAH ISSN: 0577-9189 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: Chinese (12 Refs) Under certain assumptions, the heuristic search is regarded as a random sampling process; the statistical inference method is thus introduced into the search. According to this basic idea and by incorporating the Wald sequential probability ratio test in A (or A*) search, a new heuristic search algorithm SA is obtained. The algorithm is discussed in detail and an example given, showing how it works. Desc.: computational complexity; heuristic programming; random processes; statistical analysis Ident.: random sampling process; statistical inference method; sequential probability ratio test; heuristic search algorithm Class Codes: C1230; C1140Z; C4240 Y034013 22 1588850 C86008560 The dataflow architecture: a suitable base for the implementation of expert systems Rokey, M. Dept. of Comput. Sci., Texas A&M Univ., College Station, TX, USA Comput. Archit. News (USA) vol.13, no.4 8-14 Sept. 1985 CODEN: CANED2 ISSN: 0163-5964 Treatment: GENERAL, REVIEW Doc. Type: JRNL PAP. Lang.: ENG. (8 Refs) Changes in the field of computing are inevitable. As research into the uses of computers produces more applications, conventional Von Neumann style concepts will need to be replaced with more up-to-date ideas. As an example, the field of rule-based expert systems is growing both in popularity and in the demands made on current technology. This paper examines the nature of an expert system and the characteristics which make its implementation in conventional terms unattractive. The dataflow style of computation is proposed as a viable candidate for these systems, and the current state of architectures which support this method of computation is examined. Desc.: computer architecture; expert systems; parallel processing Ident.: parallel processing; dataflow architecture; rule-based expert systems Class Codes: C5220 Y034013 23 1575262 C86005932 An expert system for control systems design LeVan, D.o. Dept. of Electr. & Comput. Eng., Buffalo Univ., NY, USA Kirk, D.E. (Editors) Sponsor: IEEE; Naval Postgraduate Sch.; Univ. Santa Clara Conference Record Eighteenth Asilomar Conference on Circuits, Systems and Computers (Cat. No. 85CH2200-4) 252-6 1985 5-7 Nov. 1984 Pacific Grove, CA, USA Publ: IEEE Comput. Soc. Press, Washington, DC, USA x+456 pp. ISBN 0 8186 0673 8 U. S. Copyright Clearance Center Code: CH2200-4/85/0000-0252$01.00 Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (6 Refs) An expert system for control system design has data representation needs not required by other common expert systems. Directed energy flows, procedures to transform energy and information, and a recursive definition are necessary. The concept of frames is modified to form an extended frame definition. This provides important items such as slots for default/required information, procedures, nested frame definitions, and ports for directing the flow of energy or information. Relationships between the slots, procedures, and frames can be defined to indicate how they all interact and form a whole. Extended frames also support the concepts and constructs needed for use by an inference engine. Desc.: control system CAD; expert systems Ident.: directed energy flows; knowledge-based system; CAD; computer-aided design; expert system; control systems design; recursive definition; extended frame definition; inference engine Class Codes: C7420; C1310 Y034013 24 1575257 C86006005 An expert planning system for air-to-air combat management Markosian, L.Z.; Rockmore, A.J.; Keller, K.J.; Gaan, M.R. Syst. Control Technol. Inc., Palo Alto, CA, USA Kirk, D.E. (Editors) Sponsor: IEEE; Naval Postgraduate Sch.; Univ. Santa Clara Conference Record Eighteenth Asilomar Conference on Circuits, Systems and Computers (Cat. No. 85CH2200-4) 112-16 1985 5-7 Nov. 1984 Pacific Grove, CA, USA Publ: IEEE Comput. Soc. Press, Washington, DC, USA x+456 pp. ISBN 0 8186 0673 8 U. S. Copyright Clearance Center Code: CH2200-4/85/0000-0112$01.00 Treatment: THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (6 Refs) In the past, AI planning systems have focused primarily on logical representation of the planning problem and search of the plan space. More recently, planning systems are being applied to real-world problems characterized by dynamic domains with mathematically complex representations, uncertainty of current and future domain states, and unresponsiveness to the planner's actions. A paradigm for dynamic plan execution and monitoring is presented that complements skeletal planning based on scripts. This paradigm interleaves plan generation and plan interpretation phases. Plan generation and interpretation are implemented as coroutines in an application to the highly dynamic air battle environment. In this application, the planning system combines expert fighter pilot heuristics with detailed mathematical functions that evaluate and execute these heuristics. Desc.: expert systems; military systems Ident.: expert planning system; air-to-air combat management; AI; dynamic domains; domain states; paradigm; air battle environment; fighter pilot heuristics Class Codes: C7460; C1230; C3360L Y034013 25 1573768 C86000468 A heuristic method for searching the global maximum of a multimodal function Kamei, K.; Inoue, K. Dept. of Electr. Eng., Ritsumeikan Univ., Kyoto, Japan Gertler, J.; Keviczky, L. (Editors) Sponsor: IFAC; IMACS; IFIP; IFORS; IMEKO; UNESCO; UNIDO Bridge Between Control Science and Technology. Proceedings of the Ninth Triennial World Congress of IFAC 1609-14 vol.3 1985 2-6 July 1984 Budapest, Hungary Publ: Pergamon Press, Oxford, England 6 vol. (xiv+xiii+xiv+xiv+xiv+xiv+3526) pp. ISBN 0 08 031666 2 Treatment: THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (15 Refs) Many techniques for searching the optimum of a unimodal function are discussed in nonlinear optimization problems and mathematical programming. On the other hand, though the search method for the global optimum of a multimodal function is an important technique in the optimization of systems, it is not discussed so sufficiently as the unimodal search methods. This paper describes a new method for searching the global maximum of a multimodal function. The method is composed of three kinds of searches. They are called a G (Grasping)-mode, an F (Finding)-mode and a C (Confirming)-mode searches. In the G-mode and the C-mode searches, a heuristic technique which was extracted through the search behaviour evolved by human subjects is used. In the F-mode search, the simplex method which has been well-known as a search method for a unimodal function is used. The numerical results, using a variety of test problems, are given. Desc.: heuristic programming; optimisation Ident.: grasping; finding; confirming; global maximum; nonlinear optimization; mathematical programming; G-mode; C-mode; F-mode Class Codes: C1180; C1230 Y034013 26 1567995 C86000042 Automated product development-the 'design to product' Alvey demonstrator Burrow, L.D. Electrotechnology (GB) vol.13, no.4 130-2 Oct. 1985 CODEN: ETNYBB ISSN: 0306-8552 Treatment: PRACTICAL Doc. Type: JRNL PAP. Lang.: ENG. The Alvey Directorate in the Department of Trade and Industry has established a national programme of collaborative research and development in the four areas of man-machine interfaces (MMI), intelligent knowledge based systems (IKBS), software engineering, and very large scale integration (VLSI). One of these projects is the GEC 'design to product' demonstrator. In this project current research work in IKBS, MMI, and software engineering is being applied to the automation of the complete product development route, from initial design, through manufacture to field maintenance. The core of the project is an intelligent Knowledge Based, Design System (IKBDS). This provides a 'design environment' containing general design knowledge with an interface allowing a user to use the knowledge in an efficient and flexible manner. This interaction creates a product description containing all the information required to design, manufacture and support the product. An advanced fully automated flexible manufacturing system also being developed within the project and driven by plans and NC data from the product description will demonstrate the manufacture of design components. Desc.: CAD; CAD/CAM; expert systems; flexible manufacturing systems ; research initiatives; VLSI Ident.: Alvey demonstrator; man-machine interfaces; MMI; intelligent knowledge based systems; IKBS; software engineering; very large scale integration; VLSI; design; manufacture; maintenance; automated flexible manufacturing system Class Codes: C0200; C3355; C7420 Y034013 27 1567522 C86000641 Search processes for applications of artificial intelligence Klinger, A. Dept. of Comput. Sci., California Univ., Los Angeles, CA, USA Proc. SPIE Int. Soc. Opt. Eng. (USA) vol.548 166-8 1985 CODEN: PSISDG ISSN: 0277-786X Applications of Artificial Intelligence II 9-11 April 1985 Arlington, VA, USA Treatment: APPLIC Doc. Type: CONFERENCE PAPER Lang.: ENG. (13 Refs) This paper concerns search as an aspect of artificial intelligence (AI). Search is a heavily used part of AI programs that work in the domains of problem solving, resolution theorem-proving, and speech, text, and image understanding. It is also the one out of all the AI tools, that is the most closely-coupled to practical problems by its very nature. This is because computers excel at storage of quantities of information and at repetitive execution of a set of comparisons. Of all the 'intelligent' uses of computing, search is the one realm where human solutions can be surpassed by computers. This paper begins with a brief survey of some search processes from traditional AI domains such as problem-solving. Next the author discusses the computer vision problem, and focuses on object determination. The use of recursive methods is then presented. The author explains the speed advantages of algorithms such as binary search, and how recursion is built in to them. Hierarchical methods of array search are then seen to be an ideal way to apply recursion to effectively compute topological properties without excessive increase in time needed, when the domain size expands. Hence, advantageous use of such approaches for automation, medical and remote sensing data, and other image applications of artificial intelligence is described, to illustrate the theme. Desc.: artificial intelligence; computer vision; data handling Ident.: search processes; speech understanding; text understanding; object determination; hierarchical methods; artificial intelligence; problem solving; resolution theorem-proving; image understanding; computer vision; recursive methods; binary search; array search; topological properties Class Codes: C1230; C6130 Y034013 28 1567517 C86006224 Application of expert system techniques to a visual tracker Myler, H.R.; Thompson, W.E.; Flachs, G.M. Dept. of Electr. & Comput. Eng., New Mexico State Univ., Las Cruces, NM, USA Proc. SPIE Int. Soc. Opt. Eng. (USA) vol.548 124-7 1985 CODEN: PSISDG ISSN: 0277-786X Applications of Artificial Intelligence II 9-11 April 1985 Arlington, VA, USA Treatment: APPLIC; PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (8 Refs) A structure for a visual tracking system is presented which relies on information developed from previous tracking scenarios stored in a knowledge base to enhance tracking performance. The system is comprised of a centroid tracker front end which supplies segmented image features to a data reduction algorithm which holds the reduced data in a temporary database relation. This relation is then classified via two separate modes, learn and track. Under learn mode, an external teacher-director operator provides identification and weighting cues for membership in a long-term storage relation within a knowledge base. Track mode operates autonomously from the learn mode where the system determines feature validity by applying fuzzy set membership criteria to previously stored track information in the database. Results determined from the classification generate tracker directives which either enhance or permit current tracking to continue or cause the tracker to search for alternate targets based upon analysis of a global target tracking list. The classification algorithm is based on correlative analysis of the tracker's segmented output presentation after low pass filtering derives lower order harmonics of the feature. The fuzzy set membership criteria is based on size, rotation, frame location, and past history of the feature. The first three factors are linear operations on the spectra, while the last is generated as a context relation in the knowledge base. The context relation interlinks data between features to facilitate tracker operation during feature occlusion or presence of countermeasures. Desc.: computerised pattern recognition; data reduction; expert systems; fuzzy set theory; tracking Ident.: track mode; expert system techniques; visual tracking system ; knowledge base; centroid tracker front end; segmented image features; data reduction algorithm; database relation; learn mode; external teacher-director operator; feature validity; fuzzy set membership criteria tracker directives; global target tracking list; classification algorithm; correlative analysis; low pass filtering; lower order harmonics; context relation; feature occlusion; countermeasures Class Codes: C7890; C5530; C1230; C1250 Y034013 29 1567505 C86000703 Goal-directed textured-image segmentation Laws, K.I. Artificial intelligence Center, SRI Int., Menlo Park, CA, USA Proc. SPIE Int. Soc. Opt. Eng. (USA) vol.548 19-26 1985 CODEN: PSISDG ISSN: 0277-786X Applications of Artificial Intelligence II 9-11 April 1985 Arlington, VA, USA Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (11 Refs) The SLICE segmentation system identifies image regions that differ in gray-level distribution, color, spatial texture, or some other local property. This report concentrates on textured-image segmentation using local texture-energy measures and user-delimited training regions. Knowledge of target textures or signatures is combined with knowledge of background textures by using histogram-similarity transforms to identify regions that are similar to a target texture and dissimilar to other textures. Desc.: artificial intelligence; computerised picture processing Ident.: goal directed textured image segmentation; SLICE segmentation system; gray-level distribution; color; spatial texture; local texture-energy measures; user-delimited training regions; signatures; histogram-similarity transforms; target texture Class Codes: C1250; C1230 Y034013 30 1556539 C85051823 The use of expert-systems programming techniques for the design of lead-lag compensators James, J.R.; Frederick, D.K.; Taylor, J.H. US Mil. Acad., West Point, NY, USA Sponsor: IEE International Conference - Control 85 (Conf. Publ. No. 252) 180-5 vol.1 1985 9-11 July 1985 Cambridge, England Publ: IEE, London, England 2 vol. (xix+696) pp. Treatment: PRACTICAL; THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (13 Refs) Discusses the application of expert systems programming techniques to the design of lead-lag compensators for a single-input, single-output, continuous-time plant. A design method based on first adjusting the high-frequency response with lead and constant-gain compensators followed by adjusting the low-frequency response with lag compensators has been developed. This design heuristic is presented and its ability to achieve specifications for a range of different plants is discussed. Inference engine capabilities and extensions to a conventional analysis and design package required to implement this automatic control system design method are discussed. Desc.: compensation; control system CAD; expert systems; programming Ident.: single input single output system; control system CAD; inference engine; expert-systems; programming techniques; lead-lag compensators; continuous-time plant; design method; high-frequency response ; constant-gain compensators; low-frequency response; lag compensators; design heuristic; automatic control system design method Class Codes: C7420; C6110; C1310 Y034013 31 1556413 C85049614 Characterizing the parallelism in rule-based expert systems Douglass, R.J. Los Alamos Nat. Lab., NM, USA Gallizzi, E.L.; Elam, J.; Sprague, R.H., Jr. (Editors) Sponsor: Univ. Hawaii; Univ. Southwestern Louisiana; ACM; IEEE; Community Health Comput.; Cray Res.; IBM; Plenum Proceedings of the Eighteenth Hawaii International Conference on System Sciences 1985 47-54 vol.2 1985 2-4 Jan. 1985 Honolulu, HI, USA Publ: Hawaii Int. Conference on Syst. Sci. (Distributed by Western Periodicals Co.), Honolulu, HI, USA 3 vol. (xii+704+xx+813+ix+386) pp. Treatment: THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (34 Refs) A brief review of two classes of rule-based expert systems is presented, followed by a detailed analysis of potential sources of parallelism at the production or rule level, the subrule level (including match, select and act parallelism), and at the search level (including 'AND', 'OR', and stream parallelism). The potential amount of parallelism from each source is discussed and characterized in terms of its granularity, inherent serial constraints, efficiency, speedup, dynamic behavior, and communication volume, frequency and topology. Subrule parallelism will yield, at best, two- to tenfold speedup, and rule level parallelism will yield a modest speedup on the order of 5 to 10 times. Rule level can be combined with OR, AND, and stream parallelism in many instances to yield further parallel speedups. Desc.: expert systems; parallel processing; programming theory Ident.: parallelism; rule-based expert systems; subrule level; match ; select; act parallelism; stream parallelism; serial constraints; efficiency; speedup; dynamic behavior; communication volume; frequency; topology; OR; AND Class Codes: C4240; C1230 Y034013 32 1556126 C85051068 Computer assisted pre-operative evaluation (CAPE)-a system description Keating, H.J., III; Sims, S.C.; Balk, M.A. Dept. of Med., Wilmington Med. Center, DE, USA Cohen, G.S. (Editors) Sponsor: IEEE Proceedings of the Eighth Annual Symposium on Computer Applications in Medical Care (Cat. No. 84CH2090-9) 269-72 1984 4-7 Nov. 1984 Washington, DC, USA Publ: IEEE Comput. Soc. Press, Silver Spring, MD, USA xx+1032 pp. ISBN 0 8186 0565 0 U. S. Copyright Clearance Center Code: CH2090-9/84/0000-0269$01.00 Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (8 Refs) A computer assisted preoperative evaluation system called CAPE, has been constructed. This knowledge-based consultant system, which is designed for the general internist, provides: (1) estimation of the risk of mortality and organ system failures in patients considered for surgery, (2) specific recommendations, derived from the pertinent literature, to minimize that risk, and (3) retrieval of surgical procedure-specific considerations. Although the system seems to help optimize the patient's condition for surgery, it has not yet been subjected to a controlled trial to assess its clinical utility or user satisfaction. Desc.: expert systems; medical administrative data processing; medical diagnostic computing; patient care Ident.: medical computing; expert systems; computer assisted preoperative evaluation system; CAPE; knowledge-based consultant system; internist; risk of mortality; organ system failures; patients; surgery; literature; surgical procedure-specific considerations; clinical utility; user satisfaction Class Codes: C7140; C7330 Y034013 33 1555913 C85048418 Substantial constructive induction using layered information compression: tractable feature formation in search Rendell, L. Issued by: Univ. Illinois at Urbana-Champaign, USA; Jan. 1985 14 pp. Report No.: UIUCDCS-R-85-1198 Treatment: THEORETICAL Doc. Type: REPORT Lang.: ENG. (34 Refs) This paper addresses a problem of practical inductive inference which is more difficult than any comparable work in AI. The subject of the research is realistic constructive induction: mechanized feature formation from data represented in elementary form. A feature, attribute, or property such as 'piece advantage' in checkers is much more abstract than an elementary descriptor or primitive such as contents of a checkerboard square. Features have often been used in evaluation functions; primitives are usually too detailed for this. To create higher level features from primitives (i.e. to restructure data), a new form of clustering is used which involves layering of knowledge and invariance of utility (i.e. usefulness in attaining a goal). The scheme, which is both model- and data-driven, requires almost no background, domain-specific knowledge, but rather constructs it. The method achieves considerable generality with superior noise management and low computational complexity. Although the domains addressed are difficult, initial experimental results are encouraging. Desc.: artificial intelligence Ident.: generalisation learning; practical inductive inference; AI; constructive induction; mechanized feature formation; evaluation functions primitives; clustering; noise management; computational complexity Class Codes: C1230 Y034013 34 1553041 C85048410 Automated classification of complex errors in discrete sequential tasks Ahuja, S.B.; Reggia, J.A. Dept. of Comput. Sci., Maryland Univ., College Park, MD, USA IEEE Trans. Syst., Man & Cybern. (USA) vol.SMC-15, no.4 527-32 July-Aug. 1985 CODEN: ISYMAW ISSN: 0018-9472 U. S. Copyright Clearance Center Code: 0018-9472/85/0700-0527$01.00 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (12 Refs) Use of a heuristically guided state-space search to classify complex errors in discrete sequential processes is described, and a general method for deriving heuristic function to guide such a search is introduced. A specific heuristic function and alternate sets of elementary errors are studied through the use of an implemented program. Desc.: expert systems; heuristic programming; programming theory; state-space methods Ident.: complex error classification; SLIPS simulator; automated error classification; state-space search; discrete sequential processes; heuristic function Class Codes: C1230; C4240 Y034013 35 1552988 C85048409 Invoking inference engines in an expert system IBM Tech. Disclosure Bull. (USA) vol.27, no.12 7034-7 May 1985 CODEN: IBMTAA ISSN: 0018-8689 Treatment: NEW DEVELOPMENTS; PRACTICAL Doc. Type: JRNL PAP. Lang.: ENG. This invention relates to variable control exemplified by selectively invoking one or more inference engines in an expert system. Desc.: artificial intelligence; expert systems Ident.: selectively invoking inference engines; expert system; variable control Class Codes: C1230 Y034013 36 1552859 C85048492 Perspectives of the development of the man-machine relationship Chvalovsky, V. Vypocetni Centrum CKD Praha, Czechoslovakia Inf. Syst. (Czechoslovakia) vol.14, no.2 171-84 April 1985 CODEN: ISYSDJ ISSN: 0323-1984 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: Czech (16 Refs) Searching the way out of a certain form of stagnation, which is characteristic of the level of solving automation tasks, is connected with the need for solving the so-called weakly structured tasks typical for the decision-making processes. Automated systems for decision-making support and expert systems are of interest as a way of practical utilization of research results in the sphere of artificial intelligence. A closer analysis of the questions connected with the subject indicate that the major part of these trends in the development of automation are closely related to the development of the man-computer relationship. Desc.: artificial intelligence; human factors; man-machine systems Ident.: man-machine relationship; solving automation tasks; decision-making processes; artificial intelligence; man-computer relationship Class Codes: C1270; C1230 Y034013 37 1552059 C85048416 Models and methods for representing knowledge algebra on knowledge-manipulation frames Vol'fengagen, V.E.; Yatsuk, V.Ya. Tekh. Kibern. (USSR) vol.22, no.6 4-14 Nov.-Dec. 1984 CODEN: TEKIB8 Trans in: Eng. Cybern. (USA) vol.22, no.6 79-88 Nov.-Dec. 1984 CODEN: ENCYAF ISSN: 0013-788X U. S. Copyright Clearance Center Code: 0013-788X/84/0006-0079$7.50/0 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (12 Refs) A method for representing knowledge about specific features of operation and control of dynamic objects by means of frames and networks of frames is examined. Particular types of frames are recommended. Construction of frame networks is realized by calling on the formulated algebraic tool. A production system in which sequents are used as products is developed for automating decision-making processes. Types of subject-area postulates, which can be used to shorten the search for a solution in the state space are developed. Desc.: formal languages; knowledge engineering; state-space methods Ident.: intelligent decision making; knowledge engineering; knowledge algebra; knowledge-manipulation frames; dynamic objects; frame networks; state space Class Codes: C1230; C4210 Y034013 38 1537190 C85047294 Rule refinement (for medical expert systems) Rada, R.; Rhine, Y.; Smallwood, J. Dept. of Comput. Sci., Wayne State Univ., Detroit, MI, USA Cohen, G.S. (Editors) Sponsor: IEEE Proceedings of the Eighth Annual Symposium on Computer Applications in Medical Care (Cat. No. 84CH2090-9) 62-5 1984 4-7 Nov. 1984 Washington, DC, USA Publ: IEEE Comput. Soc. Press, Silver Spring, MD, USA xx+1032 pp. ISBN 0 8186 0565 0 U. S. Copyright Clearance Center Code: CH2090-9/84/0000-0062$01.00 Treatment: EXPERIMENTAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (14 Refs) Probabilistic, rule-based expert systems are examined both in the abstract and in a program to interpret computerized tomograms of the brain. A computer program has been written and tested that reads rules of an expert system and a learning set and automatically adjusts weights on rules so as to improve performance of the expert system. The experiments on the genetic and heuristic algorithm address more complex problems of rule adjustment. The genetic algorithm has not proven itself particular useful in these experiments but the general importance of continuity in search spaces has been demonstrated. Adjustment of weights on rules does seem effective when careful credit assignment and small weight changes are made. Adjusting the predicates in rules required, in addition, a relationship-rich store of domain-specific knowledge. Desc.: brain; computerised tomography; expert systems; heuristic programming; medical computing Ident.: rule refinement probabilistic rule based expert systems; search space continuity; medical expert systems; computerized tomograms; brain; learning set; weights; heuristic algorithm; rule adjustment; genetic algorithm; credit assignment; predicates; domain-specific knowledge Class Codes: C7330; C1230; C7140 Y034013 39 1532882 C85043861 Optimal and suboptimal feature selection for classification of evoked brain potentials Halliday, D.L.; McGillem, C.D.; Westerkamp, J.; Aunon, J.I. Bendex Guidance Syst. Div., Mishawaka, IN, USA IEEE Trans. Syst., Man & Cybern. (USA) vol.SMC-15, no.3 442-8 May-June 1985 CODEN: ISYMAW ISSN: 0018-9472 U. S. Copyright Clearance Center Code: 0018-9472/85/0500-442$01.00 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (14 Refs) Exhaustive feature selection algorithms are optimal because all possible combinations of features are tested against a predetermined criterion. Suboptimal algorithms that trade performance for speed by considering only a subset of all feature combinations are generally preferred. An implementation of the exhaustive search feature selection (ESFS) method is described for the Bayes Gaussian statistics. The algorithm significantly reduces the computational and time requirements normally associated with optimal algorithms. The performance of this algorithm is compared to that of two suboptimal algorithms-forward sequential features selection and stepwise linear discriminant analysis. Results show that this implementation provides a moderate improvement in classification accuracy and is well suited for evaluating the performance of suboptimal algorithms. Desc.: algorithm theory; Bayes methods; brain models; pattern recognition Ident.: statistical pattern recognition; evoked brain potentials; exhaustive search feature selection; Bayes Gaussian statistics; optimal algorithms; forward sequential features selection; stepwise linear discriminant analysis; classification accuracy; suboptimal algorithms Class Codes: C1250; C4240; C1230; C1140Z Y034013 40 1532861 C85043792 A subdivision algorithm in configuration space for findpath with rotation Brooks, R.A.; Lozano-Perez, T. Artificial Intelligence Lab., MIT, Cambridge, MA, USA IEEE Trans. Syst., Man & Cybern. (USA) vol.SMC-15, no.2 224-33 March-April 1985 CODEN: ISYMAW ISSN: 0018-9472 U. S. Copyright Clearance Center Code: 0018-9472/85/0300-0224$01.00 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (11 Refs) A recursive cellular representation for configuration space is presented along with an algorithm for searching that space for collision-free paths. The details of the algorithm are presented for polygonal obstacles and a moving object with two translational and one rotational degrees of freedom. Desc.: artificial intelligence; position control; robots Ident.: robotics; artificial intelligence; subdivision algorithm; findpath; rotation; recursive cellular representation; configuration space collision-free paths; polygonal obstacles Class Codes: C1230; C3120C Y034013 41 1532470 C85046619 Resident expert (expert systems) Home Comput. Adv. Course (GB) no.66 1301-3 1985 CODEN: HCACEZ ISSN: 0265-2919 Treatment: GENERAL, REVIEW Doc. Type: JRNL PAP. Lang.: ENG. For previous pt. see ibid., no.65, p.1281-3, 1985. This instalment of the series on artificial intelligence focuses on expert systems-highly structured programs capable of fulfilling much of the diagnostic and advisory role of professional human practitioners in specialised areas of knowledge. An attempt is made to isolate the essential ingredients of the true expert system. It is concluded that the knowledge base, the inference engine, the knowledge-acquisition module and the explanatory system are essential; these are each discussed in turn. Desc.: expert systems Ident.: artificial intelligence; expert systems; knowledge base; inference engine; knowledge-acquisition module; explanatory system Class Codes: C7000 Y034013 42 1532457 C85047237 Preceptor: a shell for medical expert systems and its application in a study of prognostic indices in stroke McSherry, D.; Fullerton, K. Queen's Univ. of Belfast, N Ireland Expert Syst. (GB) vol.2, no.3 140-7 July 1985 CODEN: EXSYEX ISSN: 0266-4720 Treatment: PRACTICAL Doc. Type: JRNL PAP. Lang.: ENG. (7 Refs) Preceptor is a shell system for building rule-based expert systems to support clinical decision-making. Its highly-structured knowledge base is designed to reflect the variety of decision-making activities in patient management. The authors describe Preceptor's knowledge base and dialogue structures, its inference strategies, and its comprehensive facilities for knowledge base creation and management. They report its application in a study of prognostic indices in stroke, and describe an approach to knowledge acquisition in this application. Desc.: expert systems; medical diagnostic computing Ident.: Preceptor; medical expert systems; prognostic indices; shell system; rule-based expert systems; clinical decision-making; knowledge base ; patient management; dialogue structures; inference strategies; stroke; knowledge acquisition Class Codes: C7330; C1230 Y034013 43 1531640 C85046616 Products bring expertise to industry, military (expert systems) Naughton, J. Dec User (GB) 29-35 July 1985 CODEN: DECUEZ ISSN: 0263-6530 Treatment: PRACTICAL Doc. Type: JRNL PAP. Lang.: ENG. The author describes how AI research has assisted in the development of expert systems. He outlines the three distinct components of an expert system (knowledge base, inference engine and user-interface) before looking at expert system applications. Desc.: artificial intelligence; expert systems; research initiatives Ident.: AI research; expert systems; knowledge base; inference engine; user-interface; applications Class Codes: C7000; C1230 Y034013 44 1531389 C85046227 The pell-mell rush into expert systems forces integration issue Manuel, T. Electronics (USA) vol.58, no.26 54-9 1 July 1985 CODEN: ELECAD ISSN: 0013-5070 Treatment: GENERAL, REVIEW Doc. Type: JRNL PAP. Lang.: ENG. Expert system should involve inference engine, knowledge base, explanation facility and a knowledge acquisition system and they should have interfaces to many other functions of computerized system of which it is part. The importance of AI software development tools is discussed and some of the expert systems available are mentioned. LISP machines for running AI software are also mentioned and compared with microcomputers. Software-development tools for non-AI-trained users move expert systems into everyday applications. Desc.: expert systems; software engineering; software tools Ident.: expert systems; inference engine; knowledge base; explanation facility; knowledge acquisition system; AI software development tools; expert systems; LISP machines Class Codes: C6115; C7000 Y034013 45 1529231 A85110484, C85047309 NASA image-based geological expert system development project for hyperspectral image analysis Chiou, W.C., Sr. NASA Ames Res. Center, Inf. Sci. Office, Moffett Field, CA, USA Appl. Opt. (USA) vol.24, no.14 2085-91 15 July 1985 CODEN: APOPAI ISSN: 0003-6935 Treatment: APPLIC; PRACTICAL Doc. Type: JRNL PAP. Lang.: ENG. (18 Refs) The NASA image-based geological expert system was applied to analyze remotely sensed hyperspectral image data. The major objective is for geologists to identify the Earth surface mineral properties directly from the airborne and spaceborne imaging spectrometer data. With certain constraints, the author showed that the system can identify correctly different classes of mineral. It has the built-in learning paradigm to enhance the confidence factor of mineral identification. The author also incorporated a very powerful natural language system as the user-friendly front end and tested the concurrent processing efficiency of the frame-based knowledge representation in the hypercube microsupercomputer simulation. Desc.: computerised picture processing; expert systems; geology; geophysical prospecting; geophysical techniques; geophysics computing; remote sensing; spectral analysis Ident.: airborne imaging spectrometer data; multispectral remote sensing; prospecting; technique; thematic mapper; NASA image-based geological expert system; hyperspectral image analysis; remotely sensed hyperspectral image data; Earth surface mineral properties; spaceborne imaging spectrometer data; built-in learning paradigm; confidence factor; mineral identification; natural language system; user-friendly front end; concurrent processing efficiency; frame-based knowledge representation; hypercube microsupercomputer simulation Class Codes: A9385; C7340; C1230; C5260 Y034013 46 1526948 B85051898, C85042831 'Humanized' system tracks troubles (AT&T Automated Cable Expertise system) Wilkinson, P. Teleph. Eng. & Manage. (USA) vol.89, no.8 86-91 15 April 1985 CODEN: TPEMAW ISSN: 0040-263X Treatment: PRACTICAL; PRODUCT REVIEW Doc. Type: JRNL PAP. Lang.: ENG. Discusses ACE, AT&T's Automated Cable Expertise system that diagnoses and ranks problem areas in outside plant facilities. It also is designed to cut costs and improve efficiency. ACE has the distinction of being the first commercial use of Artificial Intelligence (AI) technology in a telecom product. The technology of this expert system consists of two parts, a knowledge base, formulated directly from the expertise of cable analyzers, and an inference engine, to manipulate the data as a specialized problem solver. As a result, AI enables computers to deal with problems and draw conclusions in a way which was not possible with traditional programming. ACE automates human tasks that other computer systems support. ACE results are detailed messages about specific cable problems. All ACE software is contained in a UNIX-based 3B2-300 desktop microcomputer. The system is easy to operate, using a standard CRT with optional printer. Desc.: artificial intelligence; expert systems; telecommunication systems; telecommunications computing Ident.: telecommunications computing; AT&T; UNIX; telecommunication system; Automated Cable Expertise system; ACE; Artificial Intelligence; expert system; knowledge base; cable analyzers; inference engine; computer systems; software; 3B2-300 desktop microcomputer; CRT; printer Class Codes: B6210; C7410F Y034013 47 1518934 C85042906 An expert system for control and signal processing with automatic FORTRAN code generation Gomez, C.; Quadrat, J.P.; Sulem, A.; Blankenship, G.L.; Kumar, P.; LaVigna, A.; MacEnany, D.C.; Paul, K.; Yan, I. INRIA, Le Chesnay, France; Sponsor: IEEE Proceedings of the 23rd IEEE Conference on Decision and Control (Cat. No. 84CH2093-3) 716-23 vol.2 1984 12-14 Dec. 1984 Las Vegas, NV, USA Publ: IEEE, New York, USA 3 vol. xxviii+1776 pp. U. S. Copyright Clearance Center Code: CH2093-3/84/0000-0716$01.00 Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (21 Refs) A prototype expert system for the treatment of stochastic control and nonlinear filtering problems is described with several illustrative examples. The system is written in MACSYMA and LISP. It accepts user input in natural language or symbolic form; it carries out the basic analysis of the user's problem in symbolic form; and it produces FORTRAN code for the numerical reduction of the problem. The system has a module using PROLOG to check the well-posedness of linear and nonlinear partial differential equations specified in symbolic form. Sample sessions with the system are presented to illustrate its operation. Desc.: computerised signal processing; control engineering computing; expert systems; filtering and prediction theory; stochastic systems Ident.: expert system; signal processing; FORTRAN code generation; stochastic control; nonlinear filtering; MACSYMA; LISP; natural language; PROLOG; well-posedness; partial differential equations Class Codes: C7420; C7410F; C1340G; C1260 Y034013 48 1514419 C85039730 An intelligent search for AI efficiency Durham, T. Comput. Mag. (GB) 5-6 4 July 1985 CODEN: CONPEW Treatment: GENERAL, REVIEW Doc. Type: JRNL PAP. Lang.: ENG. Search is an important concept in artificial intelligence and one that causes many problems. When AI workers say that their programs search, they mean that the program examines items that have to be generated and then tested. The search methods used are usually heuristic. The author outlines why the key to progress in AI is found by searching efficiently. Desc.: artificial intelligence; heuristic programming Ident.: artificial intelligence; search; heuristic Class Codes: C1230 Y034013 49 1503052 C85039240 Strategic planning (artificial intelligence in strategy games) Home Comput. Adv. Course (GB) no.64 1268-71 1985 CODEN: HCACEZ ISSN: 0265-2919 Treatment: APPLIC; PRACTICAL Doc. Type: JRNL PAP. Lang.: ENG. For previous pt. see ibid., no.63, p.1241-3 (1985). The application of tree search techniques in programming computers to play strategy games (such as chess, draughts and noughts and crosses) is discussed. The use of minimax techniques is explained. A BBC BASIC listing of a game in which the player tries to minimise a number whilst the computer tries to maximise it is presented as a simple illustration; a move consists of operating on the result of the opponent's move with one of four mathematical functions. Desc.: Acorn computers; artificial intelligence; BASIC listings; computer games; minimax techniques; trees (mathematics) Ident.: computer chess; artificial intelligence; strategy games; tree search techniques; chess; draughts; noughts and crosses; minimax techniques; BBC BASIC listing Class Codes: C7830D; C7000; C1160; C1180 Y034013 50 1502376 C85038279 Traditional Chinese medical expert consultation system-ZRK82 Guan Jiwen; Liu Dayou; Fu Ning; Zhu Renkang; Zhu Guanxu; Zou Mingxi; Cheng Diangin; Zhou Jixian; Liu Huanzhong Jilin Univ., China; Chin. J. Comput. (China) vol.8, no.2 112-18 1985 CODEN: JIXUDT ISSN: 0254-4164 Treatment: PRACTICAL Doc. Type: JRNL PAP. Lang.: Chinese (7 Refs) This paper presents a traditional Chinese medical expert consultation system named ZRK82. There are two objectives for this research: (1) to study and utilize the theories and the techniques of expert system in the field of artificial intelligence; (2) to investigate and develop a traditional Chinese medical expert consultation system in the dermatosis domain. This research not only emphasizes the conformity of diagnosis and treatments by the expert system to the views of the doctors but also devotes to modeling the doctors' thinking during the entire inference process. This system possesses the capabilities of being consulted by patients in their first and subsequent visits, and explaining and understanding the prescribed natural language. It also possesses some learning functions. This system runs on a Wang 2200 MVP computer with 97% accuracy in diagnosis. Desc.: expert systems; medical diagnostic computing Ident.: patient diagnosis; traditional Chinese medical expert consultation system; ZRK82; expert system; artificial intelligence; dermatosis domain; inference process; natural language; learning functions Wang 2200 MVP computer Class Codes: C7140; C7330 Y034013 51 1500699 C85035735 Purposeful arrangements via search in the version space Meystel, A. Dept. of Electr. Eng., Florida Univ., Gainesville, FL, USA Sponsor: IEEE Proceedings of the 1984 IEEE International Conference on Systems, Man and Cybernetics (Cat. No. 84CH2063-6) 112-19 1984 10-12 Oct. 1984 Halifax, NS, Canada Publ: IEEE, New York, USA 464 pp. U. S. Copyright Clearance Center Code: CH2063-6/84/0000-0112$01.00 Treatment: THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (31 Refs) A general problem of synthesizing assemblies, designing products conceptually, making system configurations, planning, etc. is considered as a problem of purposeful arrangements (PA). Most of the systems in the real world assume a description within the system of representation where any assembly can be formalized as PA. A rule-directed knowledge base is to be established that allows proper performance of the arrangement process. A recursive hierarchical search for the best arrangement should be done in the version space (constructed upon a rules-directed knowledge base), satisfying three major conditions simultaneously: matching the purpose of the arrangement formulated as an inquiry specification; self-matching providing an internal consistency; and finally the external matching with a 'meta-whole' in which the particular arrangement is only a component. Desc.: artificial intelligence Ident.: synthesizing assemblies; designing; purposeful arrangements rule-directed knowledge base; recursive hierarchical search Class Codes: C1230 Y034013 52 1500692 C85035728 Search party (searching strategies in artificial intelligence) Home Comput. Adv. Course (GB) no.63 1241-3 1985 CODEN: HCACEZ ISSN: 0265-2919 Treatment: GENERAL, REVIEW; PRACTICAL Doc. Type: JRNL PAP. Lang.: ENG. Using the example of rats in a maze, the article explains the advantages of heuristic searching over random searching and exhaustive searching as a problem-solving method. It is explained that the method is also applicable to general, nonspatial, problems (e.g. integration of symbolic mathematical expressions) and the generation of a search tree is discussed. A BASIC program listing for the BBC microcomputer is presented with notes on the changes necessary for the Sinclair Spectrum and commodore 64, the program uses the heuristic Manhattan distance method, which is explained in the article, to solve the maze search problem. Desc.: Acorn computers; artificial intelligence; BASIC listings; Commodore computers; heuristic programming; home computing; Sinclair computers; trees (mathematics) Ident.: search tree generation; artificial intelligence; rats; heuristic searching; random searching; exhaustive searching; problem-solving method; BASIC program listing; BBC microcomputer; Sinclair Spectrum; commodore 64; Manhattan distance method; maze search problem Class Codes: C1230; C7830; C7000 Y034013 53 1490297 A85077708, B85042914, C85034660 Frame-based understanding of ECG signals Lee, H.S.; Thakor, N.V. Dept. of Electr. Eng. & Comput. Sci., Northwestern Univ., Evanston, IL, USA Sponsor: IEEE; American Assoc. Artificial Intelligence First Conference on Artificial Intelligence Applications (Cat. No. 84CH2107-1) 624-9 1984 5-7 Dec. 1984 Denver, CO, USA Publ: IEEE Comput. Soc. Press, Silver Spring, MD, USA xv+665 pp. ISBN 0 8186 0624 X U. S. Copyright Clearance Center Code: CH2107-1/84/0000-0624$01.00 Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (11 Refs) The configuration of a new ECG signal understanding system is presented. The structural and semantic information regarding signals is described in frames under the name of the prototypes. A prototype-driven hypothesis formation paradigm is used in the understanding process. Extraction of the significant points of signals is followed by the understanding process. A constraint satisfaction strategy and structural pattern matching techniques are used in a mixed fashion, the former is used in building the layered graph from which globally prominent points are searched for, whereas the latter are used for confirming waves by finding the supporting points of each prominent point under study. The domain of the constraints includes time intervals and attributes of signal components. These constraints can be exploited to reduce the complexity of search effort. For recognition of QRS, an ad hoc hypothesis-switching network is maintained that facilitates dynamic QRS hypothesis switching when a hypothesized QRS would no be instantiated. Desc.: artificial intelligence; computerised pattern recognition; electrocardiography; medical computing Ident.: ECG signals; semantic information; pattern matching techniques; layered graph; signal components; hypothesis-switching network QRS hypothesis switching Class Codes: A8770E; B7210D; C7330; C1250B; C1230 Y034013 54 1490262 B85042878, C85034850 Knowledge representation in an in-circuit test program generator Hirgelt, E.S. Zehntel Autom. Syst., Walnut Creek, CA, USA Sponsor: IEEE International Test Conference 1984 Proceedings (Cat. No. 84CH2084-2) 773-7 1984 16-18 Oct. 1984 Philadelphia, PA, USA Publ: IEEE Comput. Soc. Press, Silver Spring, MD, USA xxxi+886 pp. ISBN 0 8186 0548 0 U. S. Copyright Clearance Center Code: CH2084-2/84/0000-0773$01.00 Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (2 Refs) A formal model of an in-circuit program generator based on a fact-algorithm-rule structure is described. The generalized structure of the model will allow modification of the system and its databases in directions developed by the user rather than the original designer. This approach will also provide a structure on which to base the combined in-circuit and functional test programs required in the near future. Desc.: automatic programming; automatic testing; electronic engineering computing; expert systems; integrated circuit testing Ident.: automatic testing; expert systems; IC testing; electronic engineering computing; in-circuit test program generator; model; fact-algorithm-rule structure; functional test Class Codes: B7210B; B2570; B0170E; C7410D Y034013 55 1488444 B85040875, C85034891 HAL: a heuristic approach to schematic generation Ahlstrom, M.L.; Hadden, G.D.; Stroick, G.r. Honeywell Comput. Sci. Center, Bloomington, MN, USA Sponsor: IEEE IEEE International Conference on Computer-Aided Design, ICCAD - 84. Digest of Technical Papers (Cat. No. 84CH2026-3) 83-6 1984 12-15 Nov. 1984 Santa Clara, CA, USA Publ: IEEE, New York, USA xii+288 pp. Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (12 Refs) This investigation focuses on the suitability of rule-based expert systems for the problem of schematic generation. A rule-based expert system prototype has been developed that generates aesthetically pleasing and easily understandable schematics from MIMOLA hardware descriptions at the register-transfer level. This system is implemented in OPS5 and Lisp. Although it is possible to build schematic generators with rule-based expert system techniques, even illustrative systems require a significant knowledge acquisition effort, careful system design and substantial support in the underlying programming language. Desc.: circuit layout CAD; VLSI Ident.: VLSI; HAL; heuristic approach; schematic generation; rule-based expert systems; MIMOLA hardware descriptions; register-transfer level; OPS5; Lisp; knowledge acquisition effort; programming language Class Codes: B2570; B1130B; C7410D Y034013 56 1487253 B85039548, C85034710 MELD: an implementation of a meta-level architecture for process diagnosis Thompson, T.F.; Wojcik, R.M. Westinghouse Res. & Dev. Center, Pittsburgh, PA, USA Sponsor: IEEE; American Assoc. Artificial Intelligence First Conference on Artificial Intelligence Applications (Cat. No. 84CH2107-1) 321-30 1984 5-7 Dec. 1984 Denver, CO, USA Publ: IEEE Comput. Soc. Press, Silver Spring, MD, USA xv+665 pp. ISBN 0 8186 0624 X U. S. Copyright Clearance Center Code: CH2107-1/84/0000-0321$01.00 Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (18 Refs) A description is given of an expert system architecture called MELD (meta-level diagnosis), whose primary domain is diagnosis of electromagnetical systems. The key feature of the design is the separation of task-specific or object-level heuristic and causal rules from task-independent or metalevel rules. The metalevel rules allow the system to reason introspectively about uncertain, faulty, or incomplete evidence, and about controlling the selection and application of object-level rules. A description is also given of an implementation of the architecture in which metalevel rules are written in a language that has direct access to, and facilitates the manipulation of, object-level rules and propositions. Finally, the authors describe their experience in using the system and their efforts to emulate human troubleshooters through the development of several metalevel control paradigms and the discovery of generic diagnostic techniques that reduce search for process malfunctions. Desc.: expert systems; fault location Ident.: metalevel architecture; MELD; process diagnosis; expert system architecture; electromagnetical systems; causal rules; metalevel rules; incomplete evidence; object-level rules; human troubleshooters Class Codes: B0170E; C7400; C1230 Y034013 57 1487119 C85035165 Inference rules in a geographical information system Jungert, E. Nat. Defense Res. Inst., Linkoping, Sweden Sponsor: IEEE 1984 IEEE Workshop on Languages for Automation (Cat. No. 84CH2089-1) 32-7 1984 1-3 Nov. 1984 New Orleans, LA, USA Publ: IEEE Comput. Soc. Press, Silver Spring, MD, USA ix+235 pp. ISBN 0 8186 0611 8 U. S. Copyright Clearance Center Code: CH2089-1/84/0000-0032$01.00 Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (15 Refs) Inference rules have become a rather common means for making elaborate queries in many kinds of information systems. A problem when implementing inference systems is to speed up the process. This problem is especially apparent in geographical information systems, since the objects, in most cases, must be selected from a map on a satellite picture. This makes the inference process inefficient, because the objects are unordered. A method for the ordering of geographical objects that will be used as input to the inference process is presented. The objects are ordered in such a way that the objects that are geographically close to each other are close to each other in the final ordering. The method is integrated in a geographical information system. Those aspects of the system that are relevant to the method are also described. Desc.: expert systems; geography Ident.: geographical information system; inference systems; objects satellite picture Class Codes: C7890; C7390 Y034013 58 1487065 C85035098 Computer-aided learning tool for simulation Jagannathan, V.; Elmaghraby, A.S. Appl. Math. & Comput. Sci., Louisville Univ., KY, USA Sponsor: IEEE; Coll. CAD/CAM Consortium Proceedings of the Third Annual Workshop on Interactive Computing: CAE: Electrical Engineering Education (Cat. No. 84CH2086-7) 94-8 1984 24-26 Oct. 1984 Pittsburgh, PA, USA Publ: IEEE Comput. Soc. Press, Silver Spring, MD, USA vi+115 pp. ISBN 0 8186 0592 8 U. S. Copyright Clearance Center Code: CH2086-7/84/0000-0094$01.00 Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (10 Refs) A demonstration is presented of the application of an expert system to develop a computer-aided learning tool for simulation. The inference engine of the expert system is of a general nature. A set of rules tailored to the problem of language selection for simulation studies is described and then used to demonstrate several test cases. The user interface includes some natural language generation features and is enhanced by a logical explanation scheme. Desc.: computer aided instruction; digital simulation; expert systems; learning systems Ident.: expert system; computer-aided learning tool; simulation; inference engine; language selection; user interface; natural language generation; logical explanation scheme Class Codes: C7810C Y034013 59 1486997 C85035014 A general expert system design for diagnostic problem solving Fink, P.K.; Lusth, J.C.; Duran, J.W. Southwest Res. Inst., San Antonio, TX, USA Sponsor: IEEE Proceedings of the IEEE Workshop on Principles of Knowledge-Based Systems (Cat. No. 84CH2104-8) 45-52 1984 3-4 Dec. 1984 Denver, CO, USA Publ: IEEE Comput. Soc. Press, Silver Spring, MD, USA viii+213 pp. ISBN 0 8186 0567 7 U. S. Copyright Clearance Center Code: CH2104-8/84/0000-0045$01.00 Treatment: APPLIC; PRACTICAL; THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (20 Refs) The authors present an expert system design, called the Integrated Diagnostic Model (IDM), that integrates two sources of knowledge: a shallow, reasoning-oriented experimental knowledge base and a deep, functionally-oriented physical knowledge base. To demonstrate the IDM's usefulness in the problem area of diagnosis and repair, an implementation in the mechanical domain is described, as well as an example in the medical domain. Desc.: expert systems; heat systems; knowledge engineering; mechanical engineering computing; medical diagnostic computing Ident.: medical diagnostics; mechanical diagnostics; inference engine; heating system repair; chronic middle ear infection; expert system design; diagnostic problem solving; Integrated Diagnostic Model; reasoning-oriented experimental knowledge base; functionally-oriented physical knowledge base Class Codes: C7440; C7000; C1230; C7330 Y034013 60 1486983 C85034999 R1-Soar-an experiment in knowledge-intensive programming in a problem-solving architecture Rosenbloom, P.S.; Laird, J.E.; McDermott, J.; Newell, A. Dept. of Comput. Sci., Carnegie-Mellon Univ., Pittsburgh, PA, USA Sponsor: IEEE Proceedings of the IEEE Workshop on Principles of Knowledge-Based Systems (Cat. No. 84CH2104-8) 65-72 1984 3-4 Dec. 1984 Denver, CO, USA Publ: IEEE Comput. Soc. Press, Silver Spring, MD, USA viii+213 pp. ISBN 0 8186 0567 7 U. S. Copyright Clearance Center Code: CH2104-8/84/0000-0064$01.00 Treatment: APPLIC; THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (16 Refs) The authors present an experiment in knowledge-intensive programming within a general problem-solving production system architecture called Soar. In Soar, knowledge is encoded within a set of problem spaces, yielding a system capable of reasoning from first principles. Expertise consists of additional rules that guide complex problem-space searches and substitute for expensive problem-space operators. The resulting system uses both knowledge and search when relevant. Expertise knowledge is acquired either by having it programmed, or by a chunking mechanism that automatically learns new rules reflecting the results implicit in the knowledge of the problem spaces. The approach is demonstrated on the computer-system configuration task, the task performed by the expert system, R1. Desc.: DEC computers; expert systems; knowledge engineering Ident.: surface knowledge; control knowledge; automatic knowledge acquisition; OP55; DEC PDP-11 configuration; heuristic searches; DEC VAX-11 configurations; R1-Soar; knowledge-intensive programming; problem-solving architecture; problem-solving production system architecture; problem spaces; problem-space searches; chunking mechanism; computer-system configuration task; expert system Class Codes: C7430; C1230; C6110 Y034013 61 1486879 C85034681 Segmentation of a seismic section using a run length method and AI techniques Love, P.L.; Simaan, M.A. Gulf Res. & Dev. Co., Pittsburgh, PA, USA Sponsor: IEEE; American Assoc. Artificial Intelligence First Conference on Artificial Intelligence Applications (Cat. No. 84CH2107-1) 550-5 1984 5-7 Dec. 1984 Denver, CO, USA Publ: IEEE Comput. Soc. Press, Silver Spring, MD, USA xv+665 pp. ISBN 0 8186 0624 X U. S. Copyright Clearance Center Code: CH2107-1/84/0000-0550$01.00 Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (3 Refs) A method for segmenting stacked seismic data into zones of common signal character based on texture analysis is described. Its performance is demonstrated on a line of seismic data from the Gulf of Mexico that had been manually segmented. The segmentation method uses statistics determined by examining the run-length of seismic reflection events. The run-length method is extended, through a decision process called the RESOLVER, to incorporate heuristic rules to influence the segmentation. The algorithm that applies these constraint rules is similar to rule-based search methods commonly used in artificial intelligence applications. A comparison is made between the automatic segmentations of the section and a manual interpretation. Desc.: artificial intelligence; expert systems; geophysics computing; heuristic programming Ident.: AI techniques; stacked seismic data; common signal character ; texture analysis; segmentation method; run-length; RESOLVER; heuristic rules; rule-based search methods; artificial intelligence Class Codes: C7340; C1230 Y034013 62 1486819 C85034575 Natural language interface to the question-answering system for physicians Bolc, L.; Strzalkowski, T. Inst. for Inf., Warsaw Univ., Poland Comput. & Artif. Intell. (Czechoslovakia) vol.3, no.1 31-46 1984 ISSN: 0232-0274 Treatment: PRACTICAL Doc. Type: JRNL PAP. Lang.: ENG. (22 Refs) Attempts have been made to design a natural language question-answering system with a deductive database. The system has been adapted to cooperate with physicians in a subdomain of medicine. The main goal is to allow a non-programmer to communicate with the database without prior training or experience. The system includes a Polish language interface allowing the user to create queries in the most natural way. The focus of this paper is on how a computer should transform a real Polish medical test into a sort of formal notation so that an inference process could be applied to it. Desc.: expert systems; medical computing; natural languages Ident.: natural language question-answering system; deductive database; physicians; medicine; Polish language interface; queries; medical test; formal notation; inference process Class Codes: C7330; C1230 Y034013 63 1486620 C85034315 Planning with multiple resource constraints and an application to a naval planning problem Tate, A.; Whiter, A.M. Edinburgh Univ., Scotland Sponsor: IEEE; American Assoc. Artificial Intelligence First Conference on Artificial Intelligence Applications (Cat. No. 84CH2107-1) 410-16 1984 5-7 Dec. 1984 Denver, CO, USA Publ: IEEE Comput. Soc. Press, Silver Spring, MD, USA xv+665 pp. ISBN 0 8186 0624 X U. S. Copyright Clearance Center Code: CH2107-1/84/0000-0410$01.00 Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (12 Refs) A description is given of a multiple-resource mechanism for a hierarchical, nonlinear AI planner, NONLIN, that allows plans to be generated that satisfy resource limits and account for preferences in the resources used. This mechanism relies on a local-best-then-global-best search strategy. Duration is treated as a special resource type. An application of NONLIN to a replenishment-of-naval-vessels-at-sea planning problem is described. Desc.: artificial intelligence; military computing; scheduling Ident.: multiple resource constraints; naval planning problem; multiple-resource mechanism; nonlinear AI planner; NONLIN; replenishment-of-naval-vessels-at-sea planning problem Class Codes: C7150; C1290Z; C1230 Y034013 64 1486553 C85034237 The Graduate Course Advisor: a multi-phase rule-based expert system Valtorta, M.G.; Smith, B.T.; Loveland, D.W. Dept. of Comput. Sci., Duke Univ., Durham, NC, USA Sponsor: IEEE Proceedings of the IEEE Workshop on Principles of Knowledge-Based Systems (Cat. No. 84CH2104-8) 53-7 1984 3-4 Dec. 1984 Denver, CO, USA Publ: IEEE Comput. Soc. Press, Silver Spring, MD, USA viii+213 pp. ISBN 0 8186 0567 7 U. S. Copyright Clearance Center Code: CH2104-8/84/0000-0053$01.00 Treatment: APPLIC; PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (13 Refs) The Graduate Course Advisor (GCA) is a rule-based expert system that advises graduate students in computer science. It is implemented in PROLOG, using an inference engine modeled after MYCINs. The advising task is divided into four phases, each of which may apply the inference engine to its own rule base and invoke other procedures. Some phases are diagnostic (e.g. determining a student's needs), whereas others involve planning (e.g., planning a student's schedule). The decomposition helps to manage the complexity of the advising task, it simplifies the knowledge engineering problem, and it will enhance the clarity of explanations. However, the decomposition is tightly bound to basic design decisions; a major decoupling in the GCA followed from the assumption that best schedules are almost always chosen from the top-valued courses. Desc.: educational administrative data processing; expert systems; knowledge engineering Ident.: explanation clarity; multi-phase rule-based expert system; Graduate Course Advisor; PROLOG; inference engine; diagnostic; planning; advising; knowledge engineering Class Codes: C7110; C1230 Y034013 65 1486440 C85034115 Artificial intelligence (AI) and the man-machine interface Fahnrich, K.-P.; Hanne, K.-H.; Hoepelman, J.P. Fraunhofer Inst., IAO, Stuttgart, Germany Elektrotech. Z. ETZ (Germany) vol.106, no.7-8 346-7, 350-2 April 1985 CODEN: EELZD3 ISSN: 0170-1711 Treatment: APPLIC; PRACTICAL Doc. Type: JRNL PAP. Lang.: German (11 Refs) The article discusses possible fields of application for AI, such as expert systems and speech-processing systems, and the tools available for their realisation, such as UNIX and UNIX-related systems, LISP and PROLOG and their variants, task-oriented programming (SMALLTALK, etc.), high-performance bitmap displays, and so on. It then deals with AI methods applied to man-machine interfaces, discussing disadvantages of some current solutions and possible improvements; it introduces the idea of the knowledge-based 'users assistant' leading to an 'intelligent' and adaptive interface. Dealing with generic communication modes it discusses direct manipulation, natural language, text-sensitive and speech-sensitive natural language systems, communication using formal languages and AI-supported adaptive multimodal communication. Desc.: artificial intelligence; expert systems; LISP; man-machine systems; PROLOG; speech analysis and processing Ident.: knowledge-based users assistant; AI; man-machine interface; expert systems; speech-processing systems; UNIX; UNIX-related systems; LISP ; PROLOG; task-oriented programming; SMALLTALK; high-performance bitmap displays; adaptive interface; generic communication modes; direct manipulation; natural language; text-sensitive; speech-sensitive natural language systems; formal languages; adaptive multimodal communication Class Codes: C7000; C1270; C6140D; C1230 Y034013 66 1486431 C85034106 Designing and automating the quality assessment of a knowledge-based system: the initial Automated Academic Advisor experience Cercone, N.; Hadley, R.; Martin, F.; McFetridge, P.; Strzalkowski, T. Dept. of Comput. Sci., Simon Fraser Univ., Burnaby, BC, Canada Sponsor: IEEE Proceedings of the IEEE Workshop on Principles of Knowledge-Based Systems (Cat. No. 84CH2104-8) 193-204 1984 3-4 Dec. 1984 Denver, CO, USA Publ: IEEE Comput. Soc. Press, Silver Spring, MD, USA viii+213 pp. ISBN 0 8186 0567 7 U. S. Copyright Clearance Center Code: CH2104-8/84/0000-0193$01.00 Treatment: APPLIC Doc. Type: CONFERENCE PAPER Lang.: ENG. (31 Refs) The Automated Academic Advisor (AAA), a large practical artificial intelligence system currently under development, is introduced. Two parsers are described that were designed for use with the AAA. The ATN parser ENGRA is described with emphasis given to several enhancements to traditional ATN parsers incorporated into ENGRA. The first AAA prototype consolidates ENGRA and semantic interpreter (SI) to generate formal SQL queries for the ORACLE relational data base for a limited range of queries. The authors then discuss how the ENGRA/SI prototype was enhanced. SHADOW, a Prolog-based, English analyzer that forms the basis of the second prototype, is then described. Informal comparisons are made between the applicability of the two parsers to the AAA, and the authors' initial experience with these two prototypes is discussed. The design of an evaluation subsystem is discussed briefly. Desc.: expert systems; grammars; natural languages; quality control ; query languages; relational databases; user interfaces Ident.: semantic analyser; database query generation; quality assessment; knowledge-based system; Automated Academic Advisor; artificial intelligence system; parsers; ATN parser; ENGRA; semantic interpreter; formal SQL queries; ORACLE relational data base; SHADOW; Prolog-based; English analyzer; evaluation subsystem Class Codes: C6160D; C7430; C6140D; C7890 Y034013 67 1486250 C85033921 Selective depth-first search in Prolog Palmer, M.; McKay, D.P.; Norton, L.M.; Hirschman, L.; Freeman, M.W. SDC, Burroughs Co., Paoli, PA, USA Sponsor: IEEE; American Assoc. Artificial Intelligence First Conference on Artificial Intelligence Applications (Cat. No. 84CH2107-1) 400-3 1984 5-7 Dec. 1984 Denver, CO, USA Publ: IEEE Comput. Soc. Press, Silver Spring, MD, USA xv+665 pp. ISBN 0 8186 0624 X U. S. Copyright Clearance Center Code: CH2107-1/84/0000-0400$01.00 Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (5 Refs) One of the main criticisms of Prolog as an application language for expert systems is that its left-to-right, depth-first search strategy is too rigid, resulting in potential inefficiency or even incompleteness for a wide range of problems. It is suggested that with a proper analysis of the search space, metalevel control techniques can be used to effect selective depth-first search without losing the efficient performance of Prolog expert systems. Selective depth-first search uses heuristics to reorder rather than prune alternatives in the search space. The use of selective depth-first search in an implementation of one component, the I/O base packer, of a computer configuration expert system is discussed. An illustrative example shows how careful analysis of the search spaces involved can result in the use of similar techniques, avoiding the computational overhead associated with interpreters that incorporate general-purpose flexible search strategies. Desc.: expert systems; PROLOG Ident.: Prolog; application language; expert systems; depth-first search strategy; I/O base packer; computer configuration expert system; computational overhead; interpreters; flexible search strategies Class Codes: C6140D; C7000 Y034013 68 1485644 C85032993 Implementation of very large Prolog-based knowledge bases on data flow architectures Irani, K.B.; Shih, Y. Comput. Res. Lab., Michigan Univ., Ann Arbor, MI, USA Sponsor: IEEE; American Assoc. Artificial Intelligence First Conference on Artificial Intelligence Applications (Cat. No. 84CH2107-1) 454-9 1984 5-7 Dec. 1984 Denver, CO, USA Publ: IEEE Comput. Soc. Press, Silver Spring, MD, USA xv+665 pp. ISBN 0 8186 0624 X U. S. Copyright Clearance Center Code: CH2107-1/84/0000-0454$01.00 Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (8 Refs) A Petri net model of the Prolog-type knowledge base, called the c-graph, is given as a tool for fast rule selection in conventional machines and as a precursor to future implementation on dataflow architectures. Traversing the graph is tantamount to pattern matching or rule selection. At any time during the graph traversal a token set represents the entire content of a global database. To distinguish among different global databases the tokens are 'colored' or tagged. Both forward chaining search and Prolog's backward chaining search with backtracking are discussed in the context of the c-graph. The construction of the c-graph for a production system example based on the classic monkey-and-banana problem is presented, and a possible implementation of the c-graph on a tagged dynamic dataflow machine is outlined. Desc.: computer architecture; expert systems; graph theory Ident.: Prolog-based knowledge bases; data flow architectures; Petri net; c-graph; fast rule selection; pattern matching; rule selection; token set; forward chaining; backward chaining; backtracking; monkey-and-banana problem; tagged dynamic dataflow machine Class Codes: C5220; C1230; C1160; C7000 Y034013 69 1485525 C85032745 Macro-operators: a weak method for learning Korf, R.E. Dept. of Comput. Sci., Columbia Univ., New York, NY, USA Artif. Intell. (Netherlands) vol.26, no.1 35-77 April 1985 CODEN: AINTBB ISSN: 0004-3702 U. S. Copyright Clearance Center Code: 0004-3702/85/$3.30 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (25 Refs) Explores the idea of learning efficient strategies for solving problems by searching for macro-operators. A macro-operator, or macro for short, is simply a sequence of operators chosen from the primitive operators of a problem. The technique is particularly useful for problems with non-serializable subgoals, such as Rubik's Cube, for which other weak methods fail. Both a problem-solving program and a learning program are described in detail. The performance of these programs is analyzed in terms of the number of macros required to solve all problem instances, the length of the resulting solutions (expressed as the number of primitive moves), and the amount of time necessary to learn the macros. In addition, a theory of why the method works, and a characterization of the range of problems for which it is useful are presented. The theory introduces a new type of problem structure called operator decomposability. Finally, it is concluded that the macro technique is a new kind of weak method, a method for learning as opposed to problem solving. Desc.: artificial intelligence; learning systems; macros; programming; programming theory Ident.: weak method; learning efficient strategies; macro; problem-solving program; learning program; operator decomposability Class Codes: C4240; C6110; C1230 Y034013 70 1484833 C85031700 Qualitative reasoning with deep-level mechanism models for diagnoses of mechanism failures Pan, J.Y.-C. Fairchild Lab. for Artificial Intelligence Res., Palo Alto, CA, USA Sponsor: IEEE; American Assoc. Artificial Intelligence First Conference on Artificial Intelligence Applications (Cat. No. 84CH2107-1) 295-301 1984 5-7 Dec. 1984 Denver, CO, USA Publ: IEEE Comput. Soc. Press, Silver Spring, MD, USA xv+665 pp. ISBN 0 8186 0624 X U. S. Copyright Clearance Center Code: CH2107-1/84/0000-0295$01.00 Treatment: THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (10 Refs) Expert systems are traditionally based on the production system paradigm, which encodes experts' surface-level knowledge into production rules. However, the author also observes that a human expert compiles some domain knowledge into a concise chunk of knowledge (commonly referred to as the domain model), and uses that form of knowledge in conjunction with surface level association-based knowledge in performing his domain-specific tasks. The author explores one way of using such model-encoded expert knowledge in facilitating diagnostic tasks in a domain of engineered mechanisms. In particular, he shows that by being able to reason with the deep-level domain model in explaining failure dependencies and transient behavior of a faulty mechanism, a set of heuristically generated failure hypotheses can be refined. The key technique involved is an inference process, called predictive analysis, which is capable of reasoning with the time model of the mechanism in producing qualitative descriptions of possible transient mechanism behavior. Desc.: expert systems; knowledge engineering Ident.: expert systems; deep-level mechanism models; diagnoses; mechanism failures; production system paradigm; human expert; surface level association-based knowledge; model-encoded expert knowledge; diagnostic tasks; predictive analysis; time model; transient mechanism behavior Class Codes: C1230 Y034013 71 1484832 C85031699 A computational framework for hypothesis based reasoning and its applications to perspective analysis Mulgaonkar, P.G.; Haralick, R.M.; Shapiro, L.G. SRI Int., Menlo Park, CA, USA Sponsor: IEEE; American Assoc. Artificial Intelligence First Conference on Artificial Intelligence Applications (Cat. No. 84CH2107-1) 287-94 1984 5-7 Dec. 1984 Denver, CO, USA Publ: IEEE Comput. Soc. Press, Silver Spring, MD, USA xv+665 pp. ISBN 0 8186 0624 X U. S. Copyright Clearance Center Code: CH2107-1/84/0000-0287$01.00 Treatment: THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (17 Refs) A new hypothesis-based reasoning scheme is presented for the analysis of sensory data generated by a well-defined physical process. Hypotheses are represented by subsets of allowable attributed relational tuples describing relationships between the entities in the world. Interpretations are defined to be those hypotheses that are consistent with the structure imposed by the sensing process and compatible with the observed sensory data. Although the problem of determining the interpretation of a given sensory image is equivalent to theorem proving in general, recasting it in the above hypothesis generation form makes it equivalent to a propositional logic consistency problem with side conditions on the attributes. In this form, the problem of finding the best or maximal interpretations reduces to an NP-complete problem. The authors show how to exploit the structure of the search space of all hypotheses to significantly reduce the computational complexity of this NP-complete problem. The technique is applied to the problem of interpreting line drawings of perspective projections of polyhedral objects. Desc.: artificial intelligence; computational complexity; computer vision Ident.: AI; vision; hypothesis based reasoning; perspective analysis ; sensory data; allowable attributed relational tuples; interpretation; theorem proving; propositional logic consistency problem; NP-complete problem; computational complexity; line drawings; perspective projections; polyhedral objects Class Codes: C1230; C1250 Y034013 72 1484827 C85031694 Learning of expert systems from data Cheeseman, P. SRI Int., Menlo Park, CA, USA Sponsor: IEEE Proceedings of the IEEE Workshop on Principles of Knowledge-Based Systems (Cat. No. 84CH2104-8) 115-22 1984 3-4 Dec. 1984 Denver, CO, USA Publ: IEEE Comput. Soc. Press, Silver Spring, MD, USA viii+213 pp. ISBN 0 8186 0567 7 U. S. Copyright Clearance Center Code: CH2104-8/84/0000-0115$01.00 Treatment: THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (15 Refs) A method is described for extracting information from data to form the knowledge base for a probabilistic expert system. The information that the method finds consists of joint probabilities that show significant probabilistic connections between the associated attribute values. These joint probabilities can be combined with information about particular cases to compute particular conditional probabilities. The search procedure and significance test required are presented for different types of data. The significance test requires finding if the minimum message length (in the information theory sense) required to encode the data is reduced if the joint probability being tested is given explicitly. This significance test is derived from Bayes theorem and is shown to find the hypothesis (i.e. set of significant joint probabilities) with the highest posterior probability given the data. Desc.: Bayes methods; expert systems; knowledge engineering; probability; statistical analysis Ident.: knowledge base; probabilistic expert system; search procedure; significance test; minimum message length; information theory; Bayes theorem Class Codes: C1230; C1140Z Y034013 73 1484825 C85031692 Separating and integrating control in a rule-based tool Erman, L.D.; Scott, A.C.; London, P.E. Teknowledge Inc., Palo Alto, CA, USA Sponsor: IEEE Proceedings of the IEEE Workshop on Principles of Knowledge-Based Systems (Cat. No. 84CH2104-8) 37-43 1984 3-4 Dec. 1984 Denver, CO, USA Publ: IEEE Comput. Soc. Press, Silver Spring, MD, USA viii+213 pp. ISBN 0 8186 0567 7 U. S. Copyright Clearance Center Code: CH2104-8/84/0000-0037$01.00 Treatment: PRACTICAL; THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (9 Refs) For practical expert systems, more task-specific control is needed than is provided in the built-in inference engine provided by most rule-based tools. A knowledge engineer typically accomplishes this by encoding the control within the rules or by programming outside the tool. This problem is addressed within the S.1 tool by providing a separate procedural representation for control knowledge that is integrated within the tool; this allows for a cleaner knowledge base and support by the tool for utilities such as explanation and translation of the control knowledge. Desc.: expert systems; knowledge engineering Ident.: S1; rule based backward chaining systems; integrating control; rule-based tool; expert systems; task-specific control; built-in inference engine; knowledge engineer; S.1; procedural representation; control knowledge; explanation; translation Class Codes: C1230; C7000 Y034013 74 1484820 C85031687 Automated induction proofs using methods of program synthesis Biundo, S.; Zboray, F. Inst. of Comput. Sci., Karlsruhe Univ., Germany Comput. & Artif. Intell. (Czechoslovakia) vol.3, no.6 473-81 1984 ISSN: 0232-0274 Treatment: PRACTICAL; THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (10 Refs) Presents a method to prove automatically by induction the theorems involving existentially quantified variables. The method of how to shift the burden of searching for skolem functions and their constructive definitions from the user to the system using techniques from the automated program synthesis area is shown. Desc.: automatic programming; theorem proving Ident.: automated induction proofs; program synthesis; existentially quantified variables; skolem functions Class Codes: C1230 Y034013 75 1470063 C85031017 Inductive learning of spelling-to-phoneme rules Wolff, J.G. Issued by: IBM United Kingdom Ltd., Winchester, Hants., England; Sept. 1984 ii+42 pp. Report No.: UKSC 128 Treatment: PRACTICAL Doc. Type: REPORT Lang.: ENG. (12 Refs) Describes a method of creating spelling-to-phoneme rules automatically, and a computer program in which the method is embodied. Rules are abstracted from a dictionary, in which each word is represented in orthographic and phonemic form. In the first stage of the method, a hill-climbing search is made for clusters of contiguous symbols using frequency and size of cluster as a guide to which amongst the manifold alternatives are best. Clusters are built up from the domain of letters and of phonemes and across the two domains. Every cluster which links letters with phonemes may be seen as a rule for converting a pattern of letters into a corresponding phoneme pattern (or vice versa). In many cases these rules are ambiguous: a given letter pattern may, in two or more rules, be linked to more than one phoneme pattern. These ambiguities are eliminated in the second stage of the method by allowing the letter-phoneme clusters built in the first stage to be themselves incorporated as the components of larger clusters. These large grouping are added to the rule system progressively until no more ambiguity remains. Desc.: learning systems; linguistics; speech synthesis Ident.: orthographic form; inductive learning; parsing; spelling-to-phoneme rules; computer program; dictionary; word; phonemic form; hill-climbing search; clusters; contiguous symbols; letter pattern; ambiguity Class Codes: C7820; C5585; C1230 Y034013 76 1469666 C85030292 GENIE: a generic inference engine for medical applications Sandell, H.; Bourne, J.; Shiavi, R. Dept. of Electr. & Biomed. Eng., Vanderbilt Univ., Nashville, TN, USA Semmlow, J.L.; Welkowitz, W. (Editors) Sponsor: IEEE Frontiers of Engineering and Computing in Health Care - 1984. Proceedings - Sixth Annual Conference of the IEEE Engineering in Medicine and Biology Society (Cat. No. 84CH2058-6) 66-9 1984 15-17 Sept. 1984 Los Angeles, CA, USA Publ: IEEE, New York, USA viii+857 pp. U. S. Copyright Clearance Center Code: CH2058-6/84/000-0066$01.00 Treatment: APPLIC; PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (13 Refs) GENIE is a general-purpose knowledge engineering tool for creating mixed frame- and rule-based expert systems. Its primary use has been in the creation of expert systems for evaluation of human gait. Frames are used throughout the system as a uniform data structure, with associated demons that can be triggered to perform procedural tasks as needed. Many parameters that determine system behaviour are stored in frames. Frames provide prototype structures which may be interactively pruned to build a model of the specific case studied. A frame-based agenda lists tasks to perform. Rules are collected into separate rule bases. Context-dependent variables allow rules within general knowledge to be used within specific contexts. The rule interpreter allows backward chaining, forward chaining, or data-driven rule application. Efficient control can be achieved by combining the strategies to fit the problem domain. System reasoning can be explained through display of rule and frame contents and a system for dealing with uncertainty. Desc.: expert systems; knowledge engineering; medical diagnostic computing Ident.: mixed frame expert systems; context dependent variables; efficient control; system reasoning; GENIE; generic inference engine; medical applications; general-purpose knowledge engineering tool; rule-based expert systems; human gait; uniform data structure; procedural tasks; frame-based agenda; separate rule bases; rule interpreter; backward chaining; forward chaining; data-driven rule application; problem domain; uncertainty Class Codes: C7330; C1230 Y034013 77 1469580 C85030127 Realization of a large expert problem-solving system-SYNCHEM2: a case study Gelernter, H.; Miller, G.A.; Larsen, D.L.; Berndt, D.J. Dept. of Comput. Sci., State Univ. of New York, Stony Brook, NY, USA Sponsor: IEEE; American Assoc. Artificial Intelligence First Conference on Artificial Intelligence Applications (Cat. No. 84CH2107-1) 92-106 1984 5-7 Dec. 1984 Denver, CO, USA Publ: IEEE Comput. Soc. Press, Silver Spring, MD, USA xv+665 pp. ISBN 0 8186 0624 X U. S. Copyright Clearance Center Code: CH2107-1/84/0000-0092$01.00 Treatment: APPLIC; PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (17 Refs) SYNCHEM2 is a large knowledge-based domain-specific heuristic problem-solving system that is able to find valid synthesis routes for organic molecules of substantial interest and complexity without online guidance on the part of its chemist-user. The proposed routes, which are often original, have become increasingly suggestive and useful to chemists who have experimented with the system. SYNCHEM2 draws on a very large multidimensional knowledge base containing prescriptions for executing transitions between problem states, constraints on the applicability of these prescriptions, and a collection of acceptable terminal states for the transition sequences generated through their execution. Exploration of the problem space is presently directed by a strongly tactics-oriented search algorithm, with global heuristic strategies playing a limited role in guiding the search. Desc.: chemistry computing; expert systems Ident.: chemistry; expert problem-solving system; SYNCHEM2; knowledge-based domain-specific heuristic problem-solving system; synthesis routes; organic molecules; tactics-oriented search algorithm; global heuristic strategies Class Codes: C7320 Y034013 78 1469110 C85029588 Are you an expert on the expert system? Cros, J.F. Mesures (France) vol.50, no.4 79, 81, 83-4 18 March 1985 CODEN: MSRSET ISSN: 0755-219X Treatment: GENERAL, REVIEW Doc. Type: JRNL PAP. Lang.: French Some of the 100 expert systems already operational (mostly having medical applications) are reviewed. Rules of deduction applicable to the knowledge base are explained and their sequential scanning by an inference engine (or theorem prover) is outlined. Metarules for metaknowledge are introduced, and several programming languages are mentioned. A few French companies seeking entry into the market are cited, and allusion is made to French ideas at work in the fifth-generation computer field. Desc.: expert systems Ident.: medical applications; metarules; knowledge base systems; expert systems; sequential scanning; inference engine; metaknowledge; programming languages; fifth-generation computer field Class Codes: C7000 Y034013 79 1469105 C85029583 A frame-like knowledge representation system for computer vision Numao, M.; Ishizuka, M. Inst. of Ind. Sci., Tokyo Univ., Minato-ku, Japan Sponsor: Int. Assoc. Pattern Recognition; Canadian Inf. Processing Soc.; Canadian Image Processing & Pattern Recognition Soc Seventh International Conference on Pattern Recognition (Cat. No. 84CH2046-1) 1128-30 vol.2 1984 30 July-2 Aug. 1984 Montreal, Que., Canada Publ: IEEE Comput. Soc. Press, Silver Spring, MD, USA 2 vol. (xxxviii+1385) pp. ISBN 0 8186 0545 6 U. S. Copyright Clearance Center Code: CH2046-1/84/0000-1128$01.00 Treatment: THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (6 Refs) A frame-like knowledge representation system names FBSS has been developed as a kernel of flexible computer vision systems. The FBSS provides a framelike data structure and a standard built-in search mechanism. Hierarchical model definitions required in the vision systems can be easily described by using the frames. For matching between object models and observed patterns, the built-in search mechanism can start from an arbitrary hierarchical level designated by an external program. An application to 3-D vision is illustrated. Desc.: computer vision; data structures; knowledge engineering Ident.: hierarchical model; frame-like knowledge representation system; FBSS; computer vision; data structure; search mechanism; matching; patterns; 3-D vision Class Codes: C6160Z; C1250; C1230; C5260 Y034013 80 1469096 C85029574 The design of an expert system for database design Bouzeghoub, M.; Gardarin, G. Projet SABRE, INRIA, Le Chesnay, France Gardarin, G.; Gelenbe, E. (Editors) New Applications of Data Bases. Based on Proceedings of a Workshop 203-23 1984 2-3 Sept. 1983 Cambridge, England Publ: Academic Press, London, England xii+273 pp. ISBN 0 12 275550 2 Treatment: APPLIC Doc. Type: CONFERENCE PAPER Lang.: ENG. (20 Refs) Describes the design of a specific expert system, called SECSI, which produces a relational schema with additional integrity constraints from a description of the real world given in a subset of French. The knowledge base of the expert system is composed of a semantic network portraying the application, and a set of production rules allowing the system to transform the semantic network into a relational schema. The inference engine generates the meta-base of the relational system, called SABRE. Desc.: expert systems; relational databases Ident.: database design; expert system; relational schema; integrity constraints; knowledge base; semantic network; production rules; inference engine Class Codes: C6160D Y034013 81 1466993 C85026389 Inferencing on linguistically based semantic structures Hajicova, E.; Hnatkova, M. Dept. of Appl. Math., Charles Univ., Praha, Czechoslovakia 10th International Conference on Computational Linguistics. 22nd Annual Meeting of the Association for Computational Linguistics. Proceedings of Coling 84 291-7 1984 2-6 July 1984 Stanford, CA, USA Publ: Assoc. Comput. Linguistics, Morristown, NJ, USA xv+561 pp. Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (16 Refs) The paper characterizes natural language inferencing in the TIBAQ method of question-answering, focusing on three aspects: (i) specification of the structures on which the inference rules operate, (ii) classification of the rules that have been formulated and implemented up to now, according to the kind of modification of the input structure the rules invoke, and (iii) discussion of some points in which a properly designed inference procedure may help the search of the answer, and vice versa. Desc.: computational linguistics; expert systems; natural languages Ident.: expert systems; computational linguistics; linguistically based semantic structures; natural language inferencing; TIBAQ method; question-answering; specification; structures Class Codes: C1230; C4210 Y034013 82 1466976 C85026372 A new heuristic search technique-algorithm SA Zhang, B.; Zhang, L. Dept. of Comput. Eng. & Sci., Tsinghua Univ., Beijing, China IEEE Trans. Pattern Anal. & Mach. Intell. (USA) vol.PAMI-7, no.1 103-7 Jan. 1985 CODEN: ITPIDJ ISSN: 0162-8828 U. S. Copyright Clearance Center Code: 0162-8828/85/0100-0103$01.00 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (13 Refs) The authors present a new heuristic searching algorithm by introducing the statistical inference method on the basis of algorithm A (or A*). It is called algorithm SA. In a simplified search space, a uniform m-ary tree, the authors obtain the following result: using algorithm SA, a goal node can be found with probability, one, and its mean complexity is O (N*lnN) where N is the depth at the which the goal is located. Desc.: artificial intelligence; statistical analysis Ident.: computational complexity; heuristic search; statistical inference method; goal node; probability Class Codes: C1230; C1140Z Y034013 83 1466963 C85026359 Learning in parallel networks Hinton, G.E. Dept. of Comput. Sci., Carnegie-Mellon Univ., Pittsburgh, PA, USA BYTE (USA) vol.10, no.4 265-73 April 1985 CODEN: BYTEDJ ISSN: 0360-5280 U. S. Copyright Clearance Center Code: 0360-5280/83$1.50 Treatment: GENERAL, REVIEW; THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (3 Refs) It is easy to speculate that the brain uses quite different computational principles from those used by computers but it is hard to discover what those principles are. Empirical studies of the behavior of single neurons and their patterns of connectivity have revealed many interesting facts, but the underlying computational principles are still unclear. We don't know, for example, how the brain represents complex ideas, how it searches for good matches between stored models of objects and the incoming sensory data, or how it learns. The author describes one old and one new theory of how learning could occur in brain-like parallel networks. Desc.: artificial intelligence; computer networks; learning systems ; parallel processing Ident.: artificial intelligence; computational principles; neurons; connectivity; complex ideas; stored models; sensory data; learning; brain-like parallel networks Class Codes: C1230; C5620; C1240; C5220 Y034013 84 1451874 C85025055 GUMMEX-AN EXPERT SYSTEM FOR GENERATING PROCESS-PLANS IUDICA, N.R. BATTELLE INST., FRANKFURT-MAIN, GERMANY NACHR. DOK. (GERMANY) VOL.36, NO.1 22-7 FEB. 1985 CODEN: NADOAW ISSN: 0027-7436 Treatment: PRACTICAL Doc. Type: JRNL PAP. Lang.: GERMAN (6 Refs) THIS PAPER DESCRIBES GUMMEX AN EXPERT SYSTEM FOR GENERATING PROCESS-PLAN FOR RUBBER PRODUCTS, WHICH PRODUCES A DOCUMENT FOR THE SHOP FLOOR DETAILING THE SEQUENCE OF MANUFACTURING OPERATIONS, THE MACHINES INVOLVED AND THE TECHNICAL DETAILS OF THE MANUFACTURING PROCESS SUCH AS TIMES TEMPERATURES, WEIGHTS, DIMENSIONS, ETC. GUMMEX CONSISTS OF A GRAPHICS PACKAGE, A SET OF ANCILLARY PROGRAMS, AN INFERENCE ENGINE AND THREE RULE-MODULES. THE GRAPHICS PACKAGE ALLOWS THE INTERACTIVE DESCRIPTION OF PARTS BELONGING TO PREDEFINED FAMILIES PRODUCING THE CORRESPONDING DRAWINGS. THE PLANNING MECHANISM OF GUMMEX IS IMPLEMENTED AS FORWARD-CHAINING PRODUCTION SYSTEMS, ITS KNOWLEDGE BASE COMPRISES AT PRESENT ABOUT FOUR HUNDRED PRODUCTION RULES . Desc.: EXPERT SYSTEMS; PRODUCTION CONTROL; MANUFACTURING DATA PROCESSING; RUBBER INDUSTRY Ident.: EXPERT SYSTEM; PROCESS-PLANS; GUMMEX; RUBBER PRODUCTS; SHOP FLOOR; MANUFACTURING OPERATIONS; TEMPERATURES; WEIGHTS; DIMENSIONS; GRAPHICS PACKAGE; INFERENCE ENGINE; FORWARD-CHAINING PRODUCTION SYSTEMS Class Codes: C7160 Y034013 85 1451741 C85024889 REPRESENTING AND MANIPULATING FUZZY KNOWLEDGE ABOUT FORECASTING PROBLEMS AND TECHNIQUES WHALEN, T.; SCHOTT, B. DECISION SCI. LAB., GEORGIA STATE UNIV., ATLANTA, GA, USA Sponsor: IEEE 1983 PROCEEDINGS OF THE INTERNATIONAL CONFERENCE ON SYSTEMS, MAN AND CYBERNETICS (CAT. NO. 83CH1962-0) 127-31 VOL.1 1983 29 DEC. 1983-7 JAN. 1984 BOMBAY AND NEW DELHI, INDIA Publ: IEEE, NEW YORK, USA 2 VOL. (XXI+1267+X) pp. U. S. Copyright Clearance Center Code: CH1962-0/83/0000-0127$00.75 Treatment: THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (14 Refs) THE STRUCTURE OF THE MODIFIED FUZZY PRODUCTION SYSTEM THAT WAS DEVELOPED FOR THE DECISION-SUPPORT SYSTEM FINDEX IS EXAMINED. THE INFERENCE ENGINE OF THIS PRODUCTION SYSTEM COMBINES DATA-DRIVEN AND GOAL-DIRECTED ASPECT IN A HIERARCHICAL MANNER, UNLIKE PREVIOUS FUZZY PRODUCTION SYSTEMS THAT OPERATE IN A DATA-DRIVEN MODE ONLY. THE USER INTERFACES, AS WELL AS THE PRODUCTION RULES THEMSELVES, RELY ON AN EXTENSION OF F. WENSTOP'S (1980) SYSTEM FOR FUZZY LINGUISTIC PROCESSING. IMPORTANT FEATURES OF THIS EXTENSION INCLUDE A VOCABULARY TAILORED TO THE KNOWLEDGE SOURCE USED TO DERIVE THE RULES AND THE INTRODUCTION OF A NATURAL FORM OF 'LINGUISTIC LOGARITHM'. Desc.: DECISION SUPPORT SYSTEMS Ident.: FUZZY KNOWLEDGE; FORECASTING PROBLEMS; FUZZY PRODUCTION SYSTEM; DECISION-SUPPORT SYSTEM; FINDEX Class Codes: C7102 Y034013 86 1451363 C85024475 LOOKUP ROUTINES FOR FILENAMES MANSFIELD, C. COMPUT. MAG. (GB) 23 14 MARCH 1985 CODEN: CONPEW Treatment: PRACTICAL Doc. Type: JRNL PAP. Lang.: ENG. UNDER THE DEC OPERATING SYSTEM RSTS/E PROVISION HAS BEEN MADE FOR PROGRAMMERS TO ACCESS AND READ THE OS FILE PROCESSOR (FIP) DETAILS VIA SYS CALLS EMBEDDED IN THEIR BASIC-PLUS SOURCE LANGUAGE CODE. THE AUTHOR GIVES A MODULAR SUBROUTINE IN BASIC-PLUS, WHICH WILL WORK AT LEAST UNDER VERSIONS 6C, 7 AND 7.1 OF RSTS. WHEN CALLED THE SUBROUTINE WILL SEARCH THE USERS' ACCOUNT ON DISK PACK ONE FOR ALL FILES AND LIST THE FILENAMES. EXTRA INFORMATION ON THE DATE AND TIME THAT THE FILE WAS CREATED ON THE SYSTEM IS ALSO INCLUDED BY THE LOOKUP FACILITY. Desc.: FILE ORGANISATION; BASIC LISTINGS; SUBROUTINES; TABLE LOOKUP Ident.: LOOKUP ROUTINES; FILENAMES; DEC OPERATING SYSTEM; RSTS/E; ACCESS; READ; FILE PROCESSOR; SYS CALLS; SOURCE LANGUAGE CODE; SUBROUTINE; BASIC-PLUS; 6C; 7; 7.1; RSTS; SEARCH; DISK PACK; DATE; TIME Class Codes: C6130; C7000; C6120 Y034013 87 1451293 C85024398 LARGE SCALE SYSTEM SUPPORT, INTERACTION OF SOFTWARE AND HARDWARE (C/SUP 3/ SYSTEMS) FAIREY, V. Sponsor: IEE IEE COLLOQUIUM ON PROGRESS COMMAND, CONTROL AND COMMUNICATION SYSTEMS (DIGEST NO.58) 8/1-3 1983 1 JUNE 1983 LONDON, ENGLAND Publ: IEE, LONDON, ENGLAND 26 pp. Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. A NEW APPROACH IS REQUIRED IN SYSTEM ENGINEERING ASSOCIATED WITH LARGE SCALE SYSTEMS. JOINTLY, ENGINEERING AND OPERATIONAL FEATURES MUST BE ADDRESSED SO THAT AVAILABILITY IS MAXIMISED. COST MUST BE AN IMMEDIATE CONSIDERATION, BUT WHOLE LIFE COSTING WILL TEND TOWARDS EQUALITY BETWEEN EXISTING SYSTEMS AND THE 'QUANTUM' APPROACH. INDUSTRY IN CONJUNCTION WITH THE MAINTENANCE AUTHORITIES AND THE PROCUREMENT EXECUTIVE SHOULD NOW EXAMINE THIS RADICAL APPROACH TO MAINTENANCE WITH THE AIM TO PROVIDE MAXIMUM USER AVAILABILITY EMPLOYING EXPERT SYSTEMS, INTELLIGENT SOFTWARE, REDUNDANCY, CONCEPT, ENGINEERING MANAGEMENT CONTROL, AND DOCUMENTATION INFORMATION UNDER SOFTWARE CONTROL. Desc.: COMMAND AND CONTROL SYSTEMS; SYSTEMS ENGINEERING; SYSTEMS ANALYSIS Ident.: SOFTWARE-HARDWARE INTERACTION; INDUSTRY; SYSTEM SUPPORT; C/SUP 3/ SYSTEMS; SYSTEM ENGINEERING; LARGE SCALE SYSTEMS; OPERATIONAL FEATURES; WHOLE LIFE COSTING; MAINTENANCE AUTHORITIES; PROCUREMENT EXECUTIVE; MAINTENANCE; USER AVAILABILITY; EXPERT SYSTEMS; INTELLIGENT SOFTWARE; REDUNDANCY; ENGINEERING MANAGEMENT CONTROL; DOCUMENTATION INFORMATION Class Codes: C6110; C7150 Y034013 88 1450348 C85022825 SIMULATION-THE CENTRAL POINT IN THE PLANNING OF MATERIAL FLOW SYSTEMS GINTZEL, J. FRAUNHOFER INST. FUR TRANSPORTTECH. UND WAREN-DISTRIBUTION, DORTMUND, GERMANY BREITENECKER, F.; KLEINERT, W. (Editors) SIMULATIONSTECHNIK. 2. SYMPOSIUM SIMULATIONSTECHNIK (SIMULATION TECHNIQUES. 2ND SYMPOSIUM) 227-30 1984 25-27 SEPT. 1984 WIEN, AUSTRIA Publ: SPRINGER-VERLAG, BERLIN, GERMANY XII+676 pp. ISBN 3 540 13393 3 Treatment: APPLIC Doc. Type: CONFERENCE PAPER Lang.: GERMAN (4 Refs) DESCRIBES AN INTEGRATED PLANNING AND SIMULATION SYSTEM. THE PROGRAM PACKAGE WAS DESIGNED TO PROVIDE COMPUTER AID TO CLASSICAL PLANNING CYCLE, COVERING SUCH FEATURES AS: INPUT OF PLANNED TOPOLOGY OF THE LAYOUT VIA A DIGITAL PANEL, THE DATA BEING FILED IN THE COMPUTER WITH PARALLEL GRAPHIC PRESENTATION; PROJECTING THE DIRECTIONS OF MOVEMENT FOR THE TRANSPORT SYSTEM FOR THE TENTATIVE LAYOUT; IDENTIFICATION OF BOUNDARY POINTS IN RELATION TO BASIC COMPONENTS, I.E. CONVEYOR SECTIONS, SWITCHING POINTS, ETC.; DEFINITION OF ENVIRONMENTAL DATA AND CONTROL STRATEGIES AT JUNCTIONS AND POINTS; SIMULATION OF THE MODEL AND RESULTS DISPLAYED GRAPHICALLY TOGETHER WITH DETAILED RESULTS IN LISTED FORM. A FLOW CHART SHOWS THE PLANNING AND SIMULATION SYSTEM AS SEEN BY THE USER. A NUMBER OF EXTENSION MODULES ARE PROVIDED FOR THE GENERATION OF CONTROL SOFTWARE AND SIMULATION TESTING OF THE LATTER, AND FOR THE SUBSEQUENT TRAINING OF PERSONNEL. THE EFFICIENT OPERATION OF THIS TYPE OF COMPLEX PROGRAM SYSTEM DEMANDS IMPROVED METHODS AND MORE KNOWLEDGE ON THE PART OF THE USER. TO THIS END FURTHER DEVELOPMENTS ARE IN HAND BASED ON EXPERT SYSTEMS, PARTICULARLY THOSE IN USE FOR THE DIAGNOSIS OF MEDICAL CASES AND IN GEOLOGY, CHEMISTRY, ETC. WHERE ARTIFICIAL INTELLIGENCE HAS BEEN INCORPORATED, THE EXPRESSED AIM BEING TO ATTAIN A SYSTEM FOR THE AUTOMATIC OPTIMIZATION OF SIMULATION MODELS. STANDARD ALL-PURPOSE SYSTEMS, EASY TO OPERATE AND INCORPORATING DIALOGUE SYSTEMS SUCH AS THE HAM-ANS SYSTEM, ARE ALSO RECOMMENDED. Desc.: MATERIALS HANDLING; DIGITAL SIMULATION; EXPERT SYSTEMS; MANUFACTURING DATA PROCESSING Ident.: MEDICAL DIAGNOSIS; PLANNING; MATERIAL FLOW SYSTEMS; GRAPHIC PRESENTATION; TRANSPORT SYSTEM; BOUNDARY POINTS; CONVEYOR SECTIONS; SWITCHING POINTS; ENVIRONMENTAL DATA; CONTROL STRATEGIES; FLOW CHART; SIMULATION; EXTENSION MODULES; CONTROL SOFTWARE; TRAINING; EXPERT SYSTEMS; GEOLOGY; CHEMISTRY; ARTIFICIAL INTELLIGENCE; OPTIMIZATION; DIALOGUE SYSTEMS ; HAM-ANS SYSTEM Class Codes: C3320; C7420; C7160 Y034013 89 1450021 C85022428 KNOWLEDGE-BASED SYSTEMS FOR REMOTE SENSING KANAL, L.N.; LAMBIRD, D.; LAVINE, D. DEPT. OF COMPUT. SCI., MARYLAND UNIV., COLLEGE PARK, MD, USA Sponsor: IEEE 1983 PROCEEDINGS OF THE INTERNATIONAL CONFERENCE ON SYSTEMS, MAN AND CYBERNETICS (CAT. NO. 83CH1962-0) 171-5 VOL.1 1983 29 DEC. 1983-7 JAN. 1984 BOMBAY AND NEW DELHI, INDIA Publ: IEEE, NEW YORK, USA 2 VOL. (XXI+1267+X) pp. U. S. Copyright Clearance Center Code: CH1962-0/83/0000-0171$00.75 Treatment: APPLIC; THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (22 Refs) A SUMMARY IS PRESENTED OF SOME BASIC ISSUES IN THE DESIGN OF KNOWLEDGE-BASED SYSTEMS FOR UNDERSTANDING REMOTELY SENSED IMAGES, WITH PARTICULAR ATTENTION TO THE EXTRACTION OF CARTOGRAPHIC FEATURES. AMONG THE ISSUES BRIEFLY CONSIDERED ARE: THE TYPES OF KNOWLEDGE THAT MUST BE MODELED ON REMOTE SENSING EXPERT SYSTEMS; KNOWLEDGE REPRESENTATION AND KNOWLEDGE-BASED SEARCH; AND MULTISENSOR REGISTRATION, INFORMATION FUSION, AND THE RESOLUTION OF CONFLICTING EVIDENCE. ALSO OUTLINED ARE A DISTRIBUTED ARCHITECTURE AND A CONTROL STRUCTURE BASED ON A PARALLEL NONDIRECTIONAL SEARCH ALGORITHM. Desc.: EXPERT SYSTEMS; REMOTE SENSING Ident.: REMOTE SENSING; DESIGN; KNOWLEDGE-BASED SYSTEMS; CARTOGRAPHIC FEATURES; EXPERT SYSTEMS; KNOWLEDGE REPRESENTATION; KNOWLEDGE-BASED SEARCH; MULTISENSOR REGISTRATION; INFORMATION FUSION; CONFLICTING EVIDENCE Class Codes: C1290B; C1230 Y034013 90 1449869 C85022110 PATH PLANNING IN A THREE DIMENSIONAL ENVIRONMENT RUFF, R.; AHUJA, N. COORDINATED SCI. LAB., ILLINOIS UNIV., URBANA, IL, USA Sponsor: INT. ASSOC. PATTERN RECOGNITION; CANADIAN INF. PROCESSING SOC.; CANADIAN IMAGE PROCESSING AND PATTERN RECOGNITION SOC SEVENTH INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION (CAT. NO. 84CH2046-1) 188-91 VOL.1 1984 30 JULY-2 AUG. 1984 MONTREAL, QUE., CANADA Publ: IEEE COMPUT. SOC. PRESS, SILVER SPRING, MD, USA 2 VOL. (XXXVIII+1385) pp. ISBN 0 8186 0545 6 Treatment: THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (13 Refs) AN ALGORITHM FOR FINDING COLLISION-FREE TRAJECTORIES IN A THREE-DIMENSIONAL ENVIRONMENT IS PRESENTED. GIVEN THE STARTING AND DESTINATION POINTS, THE ALGORITHM CONSTRUCTS A TRAJECTORY FROM THE SEGMENTS OF THE MEDIAL AXIS OF THE FREE SPACE. THE MOVING OBJECT IS REPRESENTED BY ITS MINIMUM ENCLOSING SPHERE. THE MEDIAL AXIS IS OBTAINED BY SHRINKING THE VOXEL REPRESENTATION, AND THE TRAJECTORY IS DERIVED BY TREATING THE MEDIAL AXIS AS A GRAPH AND CONDUCTING A HEURISTIC SEARCH. BOTH PHASES ARE IMPLEMENTED IN AN INEFFICIENT MANNER IN THE CURRENT ALGORITHM. IMPROVEMENTS IN THESE AND OTHER PARTS ARE OUTLINED. Desc.: ROBOTS; POSITION CONTROL; HEURISTIC PROGRAMMING Ident.: COLLISIONFREE TRAJECTORIES; PATH PLANNING; ROBOTICS; COLLISION AVOIDANCE; THREE DIMENSIONAL ENVIRONMENT; ALGORITHM; MEDIAL AXIS MINIMUM ENCLOSING SPHERE; VOXEL REPRESENTATION; HEURISTIC SEARCH Class Codes: C1230 Y034013 91 1449865 C85022106 STRATEGIES OF SEARCH OF DERIVATION IN PROOF BASES ZHERLOV, A.K. KIBERNETIKA (USSR) VOL.20, NO.2 84-6 MARCH-APRIL 1984 CODEN: KBRNA5 ISSN: 0023-1274 Trans in: CYBERNETICS (USA) VOL.20, NO.2 278-81 MARCH-APRIL 1984 CODEN: CYBNAW ISSN: 0011-4235 U. S. Copyright Clearance Center Code: 0011-4235/84/2002-0278$08.50 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (3 Refs) CONSIDERS SOME ASPECTS OF DERIVATION IN SYSTEM OF PROOF OF THEOREMS. IN CONTRAST TO EARLIER STUDIES, THE PROOF BASES (PB) ARE ASSUMED TO BE STRUCTURED. THIS ALLOWS ONE, IN SEARCH FOR A DERIVATION, TO USE VARIOUS LINES OF BEHAVIOR-STRATEGIES AND TO COMBINE KNOWLEDGE INTO DATA FILES UNIFORM WITH RESPECT TO A PARTICULAR FEATURE. FOR STRUCTURES THAT ARE NOT TAKEN INTO ACCOUNT, IN PB, A MECHANISM OF FORMULA DEFINITION IS INTRODUCED. IN PARTICULAR, THIS ALLOWS INTERPRETING FORMULA RELATIONSHIPS OF VARIABLES AS RELATIONS, WITH THE SUBSEQUENT POSSIBLE PROOF OF THEIR PROPERTIES, SUCH AS SYMMETRY, EQUIVALENCE, CONGRUENCE WITH RESPECT TO OTHER RELATIONS, ETC., I.E. REDEFINING A STRUCTURE ON A SET. THE MECHANISM OF FUNCTIONAL DEPENDENCY ALLOWS SIMPLIFYING THE EQUALITY PROBLEM AND ALSO PERFORMING CALCULATIONS DURING THE COURSE OF DERIVATION. THE SYSTEM OF THEOREM PROOF CONSISTS OF TWO PARTS; INFORMATIONAL-PB-AND AN ALGORITHM THAT APPLIES THE TECHNIQUES OF PB TO THE FORMULAS BEING PROVED. THE OPERATION OF THE ALGORITHM ON FORMULAS IS BASED ENTIRELY ON KNOWLEDGE CONTAINED IN PB, AND THE LOGICAL OPERATIONS ARE MINIMIZED. MOST DATA FILES ARE ACCUMULATED BY THE SYSTEM COMPILERS AND, IN A SENSE, ARE CONSTANT; CALL THEM A CATALOG. SOME OF THE FILES ARE FORMED DURING THE COURSE OF PROOF. Desc.: GROUP THEORY; SET THEORY; THEOREM PROVING Ident.: PROOF BASES; STRATEGIES; FORMULA DEFINITION; FUNCTIONAL DEPENDENCY; THEOREM PROOF; LOGICAL OPERATIONS Class Codes: C1230; C1160 Y034013 92 1449861 C85022102 DISCOVERING PATTERNS IN SEQUENCES OF EVENTS DIETTERICH, T.G.; MICHALSKI, R.S. DEPT. OF COMPUT. SCI., STANFORD UNIV., STANFORD, CA, USA ARTIF. INTELL. (NETHERLANDS) VOL.25, NO.2 187-232 FEB. 1985 CODEN: AINTBB ISSN: 0004-3702 U. S. Copyright Clearance Center Code: 0004-3702/85/$3.30 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (36 Refs) GIVEN A SEQUENCE OF EVENTS (OR OBJECTS), EACH CHARACTERIZED BY A SET OF ATTRIBUTES, THE PROBLEM CONSIDERED IS TO DISCOVER A RULE CHARACTERIZING THE SEQUENCE AND ABLE TO PREDICT A PLAUSIBLE SEQUENCE CONTINUATION. THE RULE, CALLED A SEQUENCE-GENERATING RULE, IS NONDETERMINISTIC IN THE SENSE THAT IT DOES NOT NECESSARILY TELL EXACTLY WHICH EVENT MUST APPEAR NEXT IN THE SEQUENCE, BUT RATHER, DEFINES A SET OF PLAUSIBLE NEXT EVENTS. THREE BASIC RULE MODELS ARE EMPLOYED TO GUIDE THE SEARCH FOR A SEQUENCE-GENERATING RULE: DECOMPOSITION, PERIODIC, AND DISJUNCTIVE NORMAL FORM. A PROGRAM, CALLED SPARC/E, IS DESCRIBED THAT IMPLEMENTS MOST OF THE METHODOLOGY AS APPLIED TO DISCOVERING SEQUENCE GENERATING RULES IN THE CARD GAME ELEUSIS. Desc.: ARTIFICIAL INTELLIGENCE; GAME THEORY Ident.: PATTERN DISCOVERING; SEQUENTIAL EVENTS; PERIODIC FORM; SPARC/E PROGRAM; SEQUENCE-GENERATING RULE; DECOMPOSITION; DISJUNCTIVE NORMAL FORM; GAME Class Codes: C1230; C1140E Y034013 93 1433307 C85020788, D85000855 EXPERT SYSTEMS MADE SIMPLE MILL, J. COMPUT. MAG. (GB) 13-14 21 FEB. 1985 CODEN: CONPEW Treatment: APPLIC Doc. Type: JRNL PAP. Lang.: ENG. THERE HAS BEEN A LOT OF INTEREST GENERATED IN EXPERT SYSTEMS AND EXPECTATIONS OF HIGHLY SOPHISTICATED SYSTEMS REPLACING HUMAN EXPERTS IN ALL WALKS OF PROFESSIONAL LIFE. BUT THE EARLY IMPLEMENTATIONS OF EXPERT SYSTEMS WHICH ARE NOW EMERGING AIM TO REPLACE TEXTBOOKS AND MANUALS RATHER THAN HUMAN EXPERTS. BRITISH GAS IS USING ICL'S EXPERT SYSTEMS SHELL, ADVISER, TO ADVISE ON THE USE OF HERBICIDES BY BRITISH GAS ON ITS SITES. BRITISH GAS ALREADY HAD ALL THE INFORMATION ON HERBICIDES SELECTION IN REPORT FORM BUT FOUND THAT GAPS WERE FOUND IN THE KNOWLEDGE DURING THE CREATION OF THE KNOWLEDGE BASE, DUE TO GAPS IN THE REPORT. ONE OF THE MAIN DIFFICULTIES WITH EXPERT SYSTEMS AT THE MOMENT IS THE LACK OF SOFTWARE PORTABILITY, THE PROBLEM BRITISH GAS FACE IF THEY TRY TO MAKE THE SYSTEM AVAILABLE DIRECTLY TO END-USERS IN ANY OF ITS 12 REGIONS. SOME OF THESE REGIONS USE IBM SYSTEMS AND OTHERS USE SPERRY, WHEREAS ADVISER ONLY RUNS ON ICL OR DIGITAL EQUIPMENT COMPUTERS. Desc.: EXPERT SYSTEMS; ICL COMPUTERS; DEC COMPUTERS Ident.: DEC COMPUTERS; ICL COMPUTERS; EXPERT SYSTEMS; BRITISH GAS; ADVISER; INFORMATION ON HERBICIDES SELECTION; IBM SYSTEMS; DIGITAL EQUIPMENT COMPUTERS Class Codes: C7160; D2070 Y034013 94 1433070 C85020517 ARTIFICIAL INTELLIGENCE IN BASIC BURGE, U. MIKRO- AND KLEINCOMPUT. (SWITZERLAND) VOL.6, NO.6 63-6 DEC. 1984 CODEN: MKLED2 ISSN: 0251-0006 Treatment: PRACTICAL Doc. Type: JRNL PAP. Lang.: GERMAN A TECHNIQUE USED IN ARTIFICIAL INTELLIGENCE, THE STATE TREE, IS DESCRIBED AND A BASIC PROGRAM USING THE METHOD FOR SOLVING A WELL-KNOWN LOGICAL PUZZLE IS PRESENTED. Desc.: ARTIFICIAL INTELLIGENCE; BASIC LISTINGS; MICROCOMPUTER APPLICATIONS Ident.: BASIC LISTING; SEARCH TREES; ARTIFICIAL INTELLIGENCE; STATE TREE Class Codes: C7000 Y034013 95 1431523 C85018032 AN APPLICATION OF ARTIFICIAL INTELLIGENCE TO OPERATIONS RESEARCH MARCUS, R. MATH. DEPT., WESTERN WASHINGTON UNIV., BELLINGHAM, WA, USA COMMUN. ACM (USA) VOL.27, NO.10 1044-7 OCT. 1984 CODEN: CACMA2 ISSN: 0001-0782 U. S. Copyright Clearance Center Code: 0001-0782/84/1000-1044$00.75 Treatment: PRACTICAL Doc. Type: JRNL PAP. Lang.: ENG. (5 Refs) AN EXTENSION OF CRITICAL PATH SCHEDULING TO A CASE WHERE THERE ARE ALTERNATE METHODS FOR COMPLETING A PROJECT IS INTRODUCED. THREE SOLUTION METHODS ARE PRESENTED INCLUDING ONE BASED ON THE HEURISTIC SEARCH ALGORITHM USED IN ARTIFICIAL INTELLIGENCE. Desc.: SCHEDULING; CRITICAL PATH ANALYSIS; GRAPH THEORY; ARTIFICIAL INTELLIGENCE Ident.: SHORTEST-PATH PROBLEM; AI METHODS; ARTIFICIAL INTELLIGENCE; OPERATIONS RESEARCH; CRITICAL PATH SCHEDULING; HEURISTIC SEARCH ALGORITHM Class Codes: C1290F; C1230 Y034013 96 1431417 C85017859 VISUALISING ENERGY LANDSCAPES (ARTIFICIAL INTELLIGENCE) DURHAM, T. COMPUT. MAG. (GB) 22-3 31 JAN. 1985 CODEN: CONPEW Treatment: PRACTICAL Doc. Type: JRNL PAP. Lang.: ENG. CONNECTIONIST ARCHITECTURES AND SEARCH METHODS BASED ON CONTROLLED RANDOMNESS, MAY BENEFIT SPEECH RECOGNITION, MACHINE VISION AND COMPUTER DESIGN. THE AUTHOR REPORTS ON HOW THREE ARTIFICIAL INTELLIGENCE RESEARCH GROUPS HAVE BEEN TRYING TO SEE WHAT CAN BE ACHIEVED BY INTRODUCING A WAY OF WIRING A SYSTEM LIKE THE BRAIN. Desc.: ARTIFICIAL INTELLIGENCE Ident.: CONNECTIONIST ARCHITECTURES; PARALLELISM; ENERGY LANDSCAPES SEARCH METHODS; CONTROLLED RANDOMNESS; SPEECH RECOGNITION; MACHINE VISION; COMPUTER DESIGN; ARTIFICIAL INTELLIGENCE RESEARCH GROUPS; WIRING; SYSTEM; BRAIN Class Codes: C1230 Y034013 97 1431405 C85017847 A GLOBAL MINIMIZATION ALGORITHM FOR COMPUTATIONALLY INTENSIVE SIMULATION PROBLEMS MACMULLAN, S.J.; FOX, R.L.; NAUDA, A. VOGT, W.G.; MICKLE, M.H. (Editors) Sponsor: UNIV. PITTSBURGH; IEEE; ISA; SCS; INT. ASSOC. MATH. AND COMPUT. SIMULATION MODELING AND SIMULATION. VOL.14. PROCEEDINGS OF THE FOURTEENTH ANNUAL PITTSBURGH CONFERENCE 1309-15 1983 21-22 APRIL 1983 PITTSBURGH, PA, USA Publ: ISA, RESEARCH TRIANGLE PARK, NC, USA 3 VOL. 1457 pp. Treatment: THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (3 Refs) AN EFFICIENT ALGORITHM FOR GLOBAL SEARCH HAS BEEN IMPLEMENTED IN A GENERAL PURPOSE INTERACTIVE SIMULATION PROGRAM. THE RESULTING SOFTWARE SYSTEM PERMITS RAPID SET UP OF GLOBAL OPTIMIZATION PROBLEMS FOR AUTOMATIC SOLUTION BY DIGITAL COMPUTER. THE PROGRAM HAS BEEN APPLIED TO STANDARD MULTI-MODAL TEST FUNCTIONS AND TO A PRACTICAL PROBLEM IN SIGNAL PROCESSING-A FIELD IN WHICH THE NEED FOR HIGH EFFICIENCY HAS BEEN FOUND TO BE ESPECIALLY IMPORTANT BECAUSE OF THE AMOUNT OF COMPUTER TIME REQUIRED TO MAKE A PERFORMANCE EVALUATION FOR REPRESENTATIVE EXAMPLES. THE PAPER REPORTS ON THE PROGRESS MADE ON THE TESTS, AS WELL AS THE DESIGN DIFFERENCES BETWEEN THE PROGRAM AND PREVIOUSLY EXISTING PROGRAMS FOR GLOBAL SEARCH. THE PERFORMANCE OF THE GLOBAL SEARCH ALGORITHM IS COMPARED QUANTITATIVELY WITH THAT OF EXHAUSTIVE SEARCH FOR EACH OF THE TEST CASES. Desc.: DIGITAL SIMULATION; MINIMISATION Ident.: MULTIMODAL TEST FUNCTIONS; DIGITAL SIMULATION; GLOBAL MINIMIZATION; COMPUTATIONALLY INTENSIVE SIMULATION PROBLEMS; DIGITAL COMPUTER; PERFORMANCE EVALUATION Class Codes: C1220; C7000; C1180 Y034013 98 1416769 C85017273 A COMPUTER INTELLIGENT SYSTEM FOR UNDERSTANDING CHINESE HOMONYMS PAO, Y.-H.; YANG, C.-S. DEPT. OF ELECTR. ENG. AND APPL. PHYS., CASE WESTERN RESERVE UNIV., CLEVELAND, OH, USA Sponsor: IEEE; CHINESE INST. ELECTRON FIRST INTERNATIONAL CONFERENCE ON COMPUTERS AND APPLICATIONS (CAT. NO. 84CH2039-6) 841-7 1984 20-22 JUNE 1984 BEIJING, CHINA Publ: IEEE COMPUT. SOC. PRESS, SILVER SPRING, MD, USA XIV+905 pp. ISBN 0 8186 0541 3 U. S. Copyright Clearance Center Code: CH2039-6/84/0000-0841$01.00 Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (4 Refs) EACH UTTERANCE IN SPOKEN CHINESE MAY REPRESENT ANY ONE OF SEVERAL LETTERS WHICH ARE HOMONYMS OF EACH OTHER. NATIVE HUMAN LISTENERS ARRIVE AT THE CORRECT INTERPRETATION BY MAKING USE OF CONTEXT. THE AUTHORS DESCRIBE A COMPUTER-BASED INTELLIGENT INTERPRETER THAT IS CAPABLE OF CARRYING OUT THIS SAME TASK OF DISAMBIGUATION. THE KNOWLEDGE USED IS IN THE FORM OF THREE SEPARATE ENTITIES: A WORD DICTIONARY, MORPHOLOGY RULES, AND SYNTAX RULES. THE INTERPRETATION OR UNDERSTANDING IS CORRESPONDINGLY ATTAINED IN SEVERAL STEPS THROUGH LEXICAL SEARCH OF THE DICTIONARY, PARSING, AND USE OF CONTEXTUAL CONSTRAINTS. Desc.: EXPERT SYSTEMS; NATURAL LANGUAGES Ident.: EXPERT SYSTEMS; CHINESE HOMONYMS; SPOKEN CHINESE; HUMAN LISTENERS; CONTEXT; COMPUTER-BASED INTELLIGENT INTERPRETER; WORD DICTIONARY ; MORPHOLOGY RULES; SYNTAX RULES; LEXICAL SEARCH; DICTIONARY; PARSING; CONTEXTUAL CONSTRAINTS Class Codes: C7820; C1230 Y034013 99 1416585 C85017052 AUTOMATED DESIGN OF CRANKCASES: THE CARTER SYSTEM REYNIER, M.; FOUET, J.-M. LAB. DE MECANIQUE ET TECHNOL., PARIS 6 UNIV., CACHAN, FRANCE COMPUT. AIDED DES. (GB) VOL.16, NO.6 308-13 NOV. 1984 CODEN: CAIDA5 ISSN: 0010-4485 U. S. Copyright Clearance Center Code: 0010-4485/84/060308-06$03.00 Treatment: PRACTICAL Doc. Type: JRNL PAP. Lang.: ENG. (20 Refs) CARTER IS AN EXPERT SYSTEM THAT DESIGNS CRANKCASES WITHOUT ANY INTERVENTION FROM THE USER (OTHER THAN SETTING THE PROBLEM BY DESCRIBING THE MECHANISM TO BE ENCASED). IT USES SEVERAL ALGORITHMIC BLACK BOXES (HAMILTONIAN PATH IN A GRAPH, PLATE THEORY, FINITE ELEMENTS, ETC.). TO REDUCE THE COST OF THESE ALGORITHMS AS MUCH AS POSSIBLE, IT RELIES MAINLY ON THE ENGINEER'S KNOWLEDGE TO INFER SOLUTIONS, OR AT LEAST TO SET THE BOUNDARY CONDITIONS AND LIMIT THE SEARCH SPACE. SUCH A KNOWLEDGE BASE HAS BEEN SET UP STEP BY STEP WITH THE HELP OF AN ORIGINAL PDIS DEALING WITH PRODUCTIONS WRITTEN IN NATURAL LANGUAGE. Desc.: EXPERT SYSTEMS; CAD; MECHANICAL ENGINEERING COMPUTING Ident.: AUTOMATED DESIGN; MECHANICS; CAD; CRANKCASES; CARTER SYSTEM EXPERT SYSTEM; ALGORITHMIC BLACK BOXES; BOUNDARY CONDITIONS; KNOWLEDGE BASE ; PDIS; NATURAL LANGUAGE Class Codes: C7440; C7000 Y034013 100 1415875 C85016020 INTERNATIONAL CONFERENCE ON DATA ENGINEERING (CATALOG NO. 84CH2031-3) Sponsor: IEEE INTERNATIONAL CONFERENCE ON DATA ENGINEERING (CATALOG NO. 84CH2031-3) 1984 INTERNATIONAL CONFERENCE ON DATA ENGINEERING (CATALOG NO. 84CH2031-3) 24-27 APRIL 1984 LOS ANGELES, CA, USA Publ: IEEE COMPUT. SOC. PRESS, SILVER SPRING, MD, USA XIII+630 pp. ISBN 0 8186 0533 2 Treatment: PRACTICAL Doc. Type: CONFERENCE PROCEEDINGS Lang.: ENG. THE FOLLOWING TOPICS ARE DEALT WITH: ENGINEERING APPLICATIONS; ENTITY-RELATIONSHIP APPROACH; DATABASE ARCHITECTURE; USER INTERFACE LANGUAGES; STATISTICAL DATABASES; DATABASE MODELS; USER INTERFACES; QUERY OPTIMIZATION; DATABASE DESIGN; SOFTWARE QUALITY AND PRODUCTION; KNOWLEDGE BASES; CONCURRENCY CONTROL; SOFTWARE ENGINEERING; DISTRIBUTED OPERATING SYSTEMS; ARTIFICIAL INTELLIGENCE APPLICATIONS; LOCAL AREA NETWORK ENVIRONMENTS; SOFTWARE METRICS; TESTBEDS FOR DESIGN METHODOLOGY EVALUATION; PERFORMANCE ANALYSIS; IMAGE DATA AND NETWORKS. 80 PAPERS WERE PRESENTED, ALL OF WHICH ARE PUBLISHED IN FULL IN THE PRESENT PROCEEDINGS. ABSTRACT OF INDIVIDUAL PAPERS CAN BE FOUND UNDER THE RELEVANT CLASSIFICATION CODES IN THIS OR OTHER ISSUES. Desc.: ENGINEERING COMPUTING; DATABASE MANAGEMENT SYSTEMS; DATA STRUCTURES; USER INTERFACES; SOFTWARE ENGINEERING; OPERATING SYSTEMS (COMPUTERS); EXPERT SYSTEMS Ident.: ENGINEERING APPLICATIONS; ENTITY-RELATIONSHIP APPROACH; DATABASE ARCHITECTURE; USER INTERFACE LANGUAGES; STATISTICAL DATABASES; DATABASE MODELS; USER INTERFACES; QUERY OPTIMIZATION; DATABASE DESIGN; SOFTWARE QUALITY; PRODUCTION; KNOWLEDGE BASES; CONCURRENCY CONTROL; SOFTWARE ENGINEERING; DISTRIBUTED OPERATING SYSTEMS; ARTIFICIAL INTELLIGENCE; LOCAL AREA NETWORK; SOFTWARE METRICS; TESTBEDS; PERFORMANCE ANALYSIS; IMAGE DATA; NETWORKS Class Codes: C6160; C6120; C7400; C7100 Y034013 101 1415538 C85015669 AN EXPERT SYSTEM FOR THE MANIPULATION OF PROGRAMS FOUET, J.-M. LMT, CACHAN, FRANCE Sponsor: IEEE; INF. PROCESS. ASSOC. ISRAEL; ACM; CANADIAN INF. PROCESS. SOC 4TH JERUSALEM CONFERENCE ON INFORMATION TECHNOLOGY (JCIT). NEXT DECADE IN INFORMATION TECHNOLOGY (CAT. NO. 84CH2022-2) 460-7 1984 21-25 MAY 1984 JERUSALEM, ISRAEL Publ: IEEE COMPUT. SOC. PRESS, SILVER SPRING, MD, USA XII+720 pp. ISBN 0 8186 0535 9 U. S. Copyright Clearance Center Code: CH2022-2/84/0000-0460$01.00 Treatment: APPLIC; PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (12 Refs) THE SYSTEM PRESENTED IS AN AID TO DEBUGGERS OF LARGE PROGRAMS. IT RELIES ON EFFICIENT TOOLS, A GREAT DEAL OF KNOWLEDGE ABOUT PROGRAMMING TECHNIQUES (INCLUDING USUAL ERRORS AND CLUMSINESS), AND DIALOG WITH THE USER TO UNDERSTAND WHAT THE PROBLEM IS, SURROUND ITS POSSIBLE CAUSES, AND FIND THE REAL CAUSE AND CORRECT THE BUG. IT MAY ALSO MAKE BY-THE-WAY SUGGESTIONS. KNOWLEDGE IS INCREMENTALLY CONVEYED TO THE SYSTEM THROUGH NATURAL LANGUAGE PRODUCTION RULES, WHICH ARE THEN COMPILED AND BUILT INTO A PATTERN-DIRECTED INFERENCE SYSTEM. PREDICATES, VARIABLES, AND RELATIONS USED WITHIN THE RULES ARE THEMSELVES FORMALLY DESCRIBED IN A DICTIONARY, THUS ALLOWING THE RULE-COMPILER AND THE INFERENCE ENGINE TO BE TOTALLY INDEPENDENT OF THAT PARTICULAR APPLICATION. TOOLS ARE BUILT AROUND AN ORIGINAL INTERNAL REPRESENTATION OF PROGRAMS, WHICH IS INDEPENDENT OF THE NONRECURSIVE PROGRAMMING LANGUAGE, AND CAN BE ROUGHLY DESCRIBED AS A SEMANTIC TREE, ON WHICH MANY INFERENCES AND MANIPULATIONS CAN BE DONE INEXPENSIVELY. IN THE CASE OF RECURSIVE LANGUAGES, THAT REPRESENTATION CAN STILL BE USEFUL, ALTHOUGH IT IS NO LONGER A TREE. Desc.: PROGRAM DEBUGGING; DEVELOPMENT SYSTEMS; EXPERT SYSTEMS; NATURAL LANGUAGES Ident.: PROGRAM MANIPULATION; PREDICTION; EXPERT SYSTEM; DEBUGGERS; PROGRAMMING TECHNIQUES; NATURAL LANGUAGE PRODUCTION RULES; PATTERN-DIRECTED INFERENCE SYSTEM; VARIABLES; RELATIONS; DICTIONARY; RULE-COMPILER; INTERNAL REPRESENTATION; NONRECURSIVE PROGRAMMING LANGUAGE; SEMANTIC TREE; RECURSIVE LANGUAGES Class Codes: C6115; C6150G Y034013 102 1414854 C85014351 SPECIAL RELATIONS IN DEDUCTIVE PROGRAM SYNTHESIS-A SUMMARY MANNA, Z.; WALDINGER, R. DEPT. OF COMPUT. SCI., STANFORD UNIV., CA, USA Sponsor: IEEE; INF. PROCESS. ASSOC. ISRAEL; ACM; CANADIAN INF. PROCESS. SOC 4TH JERUSALEM CONFERENCE ON INFORMATION TECHNOLOGY (JCIT). NEXT DECADE IN INFORMATION TECHNOLOGY (CAT. NO. 84CH2022-2) 100-5 1984 21-25 MAY 1984 JERUSALEM, ISRAEL Publ: IEEE COMPUT. SOC. PRESS, SILVER SPRING, MD, USA XII+720 pp. ISBN 0 8186 0535 9 U. S. Copyright Clearance Center Code: CH2022-2/84/0000-0100$01.00 Treatment: THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (10 Refs) NEW HIGH-LEVEL INFERENCE RULES FOR PROGRAM SYNTHESIS ARE PRESENTED TO FACILITATE DEDUCTIONS ABOUT COMMONLY OCCURRING RELATIONS, SUCH AS EQUALITY, EQUIVALENCE, AND ORDERINGS. USING THESE RULES ALLOWS THE AUTHORS TO ABBREVIATE THEIR PROOFS AND TO OMIT TROUBLESOME AXIOMS SUCH AS TRANSITIVITY, MONOTONICITY, AND FUNCTIONAL REFLEXIVITY. THE RESULTING DERIVATIONS ARE NOT ONLY EASIER TO UNDERSTAND, BUT ALSO EASIER TO DISCOVER BECAUSE OF A REDUCED SEARCH SPACE. THE NEW RULES ARE PART OF A NONCLAUSAL DEDUCTION SYSTEM IN WHICH THE THEOREM TO BE PROVED NEED NOT BE PUT INTO CLAUSE FORM OR ANY OTHER NORMAL FORM. THE SYSTEM ALLOWS SKOLEMIZATION, I.E., THE REMOVAL OF QUANTIFIERS TO BE DEFERRED WHEN IT IS CONVENIENT TO DO SO, AS IN DEALING WITH THE EQUIVALENCE CONNECTIVE. Desc.: PROGRAMMING THEORY; THEOREM PROVING; HEURISTIC PROGRAMMING Ident.: TERM POLARITY; RELATION SUBSTITUTION; RELATION MATCHING; THEOREM PROVING; DEDUCTIVE PROGRAM SYNTHESIS; HIGH-LEVEL INFERENCE RULES; EQUALITY; EQUIVALENCE; ORDERINGS; NONCLAUSAL DEDUCTION SYSTEM; SKOLEMIZATION; EQUIVALENCE CONNECTIVE Class Codes: C4240; C1230 Y034013 103 1413772 C85012919 IMPLEMENTING PROBLEM-SOLVING STRATEGIES USING THE META-LEVEL STERLING, L. DEPT. OF ARTIFICIAL INTELLIGENCE, EDINBURGH UNIV., SCOTLAND Sponsor: IEEE; INF. PROCESS. ASSOC. ISRAEL; ACM; CANADIAN INF. PROCESS. SOC 4TH JERUSALEM CONFERENCE ON INFORMATION TECHNOLOGY (JCIT). NEXT DECADE IN INFORMATION TECHNOLOGY (CAT. NO. 84CH2022-2) 125-32 1984 21-25 MAY 1984 JERUSALEM, ISRAEL Publ: IEEE COMPUT. SOC. PRESS, SILVER SPRING, MD, USA XII+720 pp. ISBN 0 8186 0535 9 U. S. Copyright Clearance Center Code: CH2022-2/84/0000-0125$01.00 Treatment: THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (24 Refs) A METHODOLOGY IS PRESENTED THAT HAS BEEN USED SUCCESSFULLY FOR THE CONSTRUCTION OF LARGE PROBLEM-SOLVING PROGRAMS. THE KEY ISSUE ADDRESSED IS THE REPRESENTATION AND EFFECTIVE EXECUTION OF DOMAIN SPECIFIC HEURISTICS OR STRATEGIES TO GUIDE THE SEARCH FOR A SOLUTION OF THE PROBLEM. THE REPRESENTATION LANGUAGE FOR BOTH THE STRATEGIES AND THE OBJECTS OF THE PROBLEM DOMAIN IS PREDICATE CALCULUS. STRATEGIES ARE REPRESENTED BY THE AXIOMS OF A META-THEORY OF THE PROBLEM DOMAIN, AND EXECUTED BY PERFORMING INFERENCES WITHIN THIS META-THEORY. ALL INFERENCES WITHIN THE OBJECT LEVEL (THE REPRESENTATION OF THE PROBLEM DOMAIN) ARE SIMULATED USING INFERENCE WITHIN THE META-LEVEL. A SET OF PROLOG CLAUSES EXPRESSING THE STRATEGIES CAN BE CONSIDERED AS THE META-THEORY, AND PROLOG SLD-RESOLUTION CAN BE USED AS THE META-LEVEL RULE OF INFERENCE. THIS MAKES IT PARTICULARLY EASY TO ENCODE THE METHODOLOGY IN PROLOG. EXAMPLES ARE GIVEN OF APPLICATIONS TO SYMBOLIC ALGEBRA, SOLVING MECHANICS PROBLEMS, AND PROVING THEOREMS. Desc.: THEOREM PROVING; HEURISTIC PROGRAMMING Ident.: EQUATION SOLVING; PROBLEM DOMAIN META THEORY; META-LEVEL INFERENCE RULE; PROBLEM-SOLVING STRATEGIES; LARGE PROBLEM-SOLVING PROGRAMS DOMAIN SPECIFIC HEURISTICS; REPRESENTATION LANGUAGE; PREDICATE CALCULUS; OBJECT LEVEL; PROLOG CLAUSES; PROLOG SLD-RESOLUTION; SYMBOLIC ALGEBRA; PROVING THEOREMS Class Codes: C1230; C4240 Y034013 104 1413392 C85012475 'AI IS STARK NAKED FROM THE ANKLES UP' ANDERSON, I. NEW SCI. (GB) VOL.104, NO.1430 18-21 15 NOV. 1984 CODEN: NWSCAL ISSN: 0028-6664 Treatment: GENERAL, REVIEW Doc. Type: JRNL PAP. Lang.: ENG. COMPUTER SCIENTISTS IN THE US HAVE BEEN DOING SOME SOUL SEARCHING OF LATE. THERE IS GENERAL AGREEMENT, EVENT AMONG THE FAITHFUL, THAT ARTIFICIAL INTELLIGENCE (AI) HAS BEEN OVERSOLD. ALSO MANY SAY THAT THE JAPANESE EFFORT TO DEVELOP THE NEXT GENERATION OF COMPUTERS HAS BEEN MISREPRESENTED. THEIR EFFORT IS TO BE WELCOMED, NOT FEARED. AND ANOTHER EMERGING ISSUE IS THAT AI WILL PLACE COMPUTER SCIENTISTS IN MUCH THE SAME MORAL DILEMMA AS ATOMIC PHYSICISTS. THEIR WORK WILL BECOME INCREASINGLY IMPORTANT IN THE DEPLOYMENT OF NUCLEAR ARMAMENTS AND OTHER WEAPONS. A COMPUTER SCIENTIST FROM STANFORD UNIVERSITY, FOR EXAMPLE, IS SUING THE US SECRETARY OF DEFENSE. HE IS CHARGING THAT DEVELOPMENT OF THE SO-CALLED LAUNCH-ON-WARNING CAPABILITY-WHERE COMPUTERS, NOT HUMANS, COULD DECIDE TO RESPOND TO A NUCLEAR ATTACK-IS UNCONSTITUTIONAL. THE BELIEF THAT AI HAS BEEN OVERPLAYED COMES FROM THE RANKS OF COMPUTER SCIENTISTS, ACADEMICS, INDUSTRY ANALYSTS, COMPANY EXECUTIVES AND AI RESEARCHERS THEMSELVES. THERE IS A GENUINE CONCERN THAT PUBLIC EXPECTATIONS HAVE BEEN RAISED SO HIGH THAT A BACKLASH FROM INVESTORS AND FUNDING AGENCIES WILL FOLLOW. Desc.: ARTIFICIAL INTELLIGENCE Ident.: AI; NUCLEAR WEAPONS; STARK NAKED; ANKLES; ARTIFICIAL INTELLIGENCE; MORAL DILEMMA; DEPLOYMENT; NUCLEAR ARMAMENTS; LAUNCH-ON-WARNING CAPABILITY Class Codes: C0200; C1230 Y034013 105 1397080 C85011641 HAZARD: AN EXPERT SYSTEM FOR SCREENING ENVIRONMENTAL CHEMICALS ON CARCINOGENICITY GOTTINGER, H.W. BATTELLE INST., FRANKFURT, GERMANY EXPERT SYST. (GB) VOL.1, NO.2 169-76 OCT. 1984 ISSN: 0266-4720 Treatment: PRACTICAL Doc. Type: JRNL PAP. Lang.: ENG. (11 Refs) THE PURPOSE OF THIS PAPER IS TO DESIGN AN EXPERT SYSTEM OR AN 'INTELLIGENT PROCEDURE' THAT PURPORTS TO SCREEN HAZARD POTENTIALS OF ENVIRONMENTAL CHEMICALS ON THE BASIS OF STRUCTURE-ACTIVITY RELATIONSHIPS IN THE STUDY OF CHEMICAL CARCINOGENESIS, PARTICULARLY WITH RESPECT TO ANALYSING THE CURRENT STATE OF KNOWN STRUCTURAL INFORMATION ABOUT CHEMICAL CARCINOGENS AND PREDICTING THE POSSIBLE CARCINOGENICITY OF UNTESTED CHEMICALS. AN ANALYSIS OF A COMPUTERISED DATABASE OF KNOWN CARCINOGENS (KNOWLEDGE BASE) IS BEING PERFORMED USING THE STRUCTURE-ACTIVITY TREES IN ORDER TO TEST THE VALIDITY OF THE TREE AS A CLASSIFICATION SCHEME (INFERENCE ENGINE) AND TO EVALUATE TRENDS OR PATTERNS THAT MAY EXIST BETWEEN CHEMICAL STRUCTURE AND SPECIFICITY FOR TARGET TISSUE, ROUTE OF ADMINISTRATION, AND ANIMAL SPECIES. PRACTICAL APPLICATIONS OF THE STRUCTURE-ACTIVITY TREE DEPEND ON ITS EVENTUAL VALIDATION AS A PREDICTOR OF CARCINOGENIC ACTIVITY. Desc.: EXPERT SYSTEMS; MEDICAL DIAGNOSTIC COMPUTING Ident.: HAZARD; EXPERT SYSTEM; ENVIRONMENTAL CHEMICALS; CARCINOGENICITY; COMPUTERISED DATABASE; KNOWLEDGE BASE; STRUCTURE-ACTIVITY TREES; CLASSIFICATION SCHEME; INFERENCE ENGINE; CHEMICAL STRUCTURE; ANIMAL SPECIES Class Codes: C7140; C7330 Y034013 106 1395089 C85008641 COMPUTER SOFTWARE FOR INTELLIGENT SYSTEMS LENAT, D.B. SCI. AM. (USA) VOL.251, NO.3 152-60 SEPT. 1984 CODEN: SCAMAC ISSN: 0036-8733 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. THE KEY TO INTELLIGENT PROBLEM SOLVING LIES IN REDUCING THE RANDOM SEARCH FOR SOLUTIONS. TO DO SO INTELLIGENT COMPUTER PROGRAMS MUST TAP THE SAME UNDERLYING 'SOURCES OF POWER' AS HUMAN BEINGS DO. THIS ARTICLE LOOKS AT THE PRINCIPLES OF INTELLIGENT SOFTWARE. Desc.: ARTIFICIAL INTELLIGENCE; EXPERT SYSTEMS Ident.: INTELLIGENT SYSTEMS; INTELLIGENT PROBLEM SOLVING; RANDOM SEARCH; INTELLIGENT COMPUTER PROGRAMS; INTELLIGENT SOFTWARE Class Codes: C1230 Y034013 107 1381589 C85008034 TREE SEARCH BY SCOUTING LEVY, D. PRACT. COMPUT. (GB) VOL.7, NO.11 86-8 NOV. 1984 CODEN: PRCODZ ISSN: 0141-5433 Treatment: PRACTICAL Doc. Type: JRNL PAP. Lang.: ENG. DESCRIBES 'SCOUT', AN INTELLIGENT TREE SEARCHING ALGORITHM FOR USE IN GAMES PROGRAMS. THE WAY IN WHICH THE ALGORITHM WORKS IS ILLUSTRATED FROM A CHESS GAME. A FLOW CHART OF THE ALGORITHM IS INCLUDED TO ASSIST PROGRAMMERS TO INCORPORATE THE SCOUT ALGORITHM INTO GAME PLAYING PROGRAMS AS AN ALTERNATIVE TO THE CONVENTIONAL ALPHA-BETA ALGORITHMS. Desc.: GAMES OF SKILL; ARTIFICIAL INTELLIGENCE; TREES (MATHEMATICS) Ident.: SCOUT; INTELLIGENT TREE SEARCHING ALGORITHM; CHESS GAME; GAME PLAYING PROGRAMS Class Codes: C7830D; C1230; C6130 Y034013 108 1381495 C85007927 SEARCHING FOR SOFTWARE DEWEY, P.R. POP. COMPUT. (USA) VOL.3, NO.13 137-41 MID-OCT. 1984 CODEN: PRCGDB ISSN: 0194-8075 Treatment: PRACTICAL Doc. Type: JRNL PAP. Lang.: ENG. SOFTWARE DIRECTORIES HAVE MULTIPLIED ALMOST AS FAST AS THE NUMBER OF SOFTWARE TITLES. WHEN ATTEMPTING TO SELECT A DIRECTORY, YOU NEED TO ASK SEVERAL IMPORTANT QUESTIONS: WHICH DIRECTORIES ARE AVAILABLE FOR THE MACHINE YOU OWN? WHICH DIRECTORIES OFFER INDEPENDENT REVIEWS OF SOFTWARE? WHAT KIND OF INFORMATION DO YOU NEED (HOW DETAILED, ETC.), AND HOW QUICKLY DO YOU NEED IT? HOW CURRENT MUST THE INFORMATION BE, AND HOW EXHAUSTIVE? WHICH DIRECTORIES OFFER UPDATES? HOW MUCH DO THEY COST? INFORMATION IS PROVIDED ON VARIOUS SOFTWARE DIRECTORIES AVAILABLE (IN THE USA) INCLUDING: PRINTED DIRECTORIES; ONLINE AND OFFLINE DATABASES; CATALOGS; AND PUBLIC DOMAIN SOURCES. Desc.: SOFTWARE SELECTION; COMPUTER AIDED INSTRUCTION Ident.: MICROCOMPUTER SOFTWARE; SOFTWARE; SOFTWARE DIRECTORIES; PRINTED DIRECTORIES; OFFLINE DATABASES; CATALOGS; PUBLIC DOMAIN SOURCES Class Codes: C7810C; C7000; C0310 Y034013 109 1381439 C85007837 EXPERT SYSTEM FOR DIESEL ELECTRIC LOCOMOTIVE REPAIR BONISSONE, P.P.; JOHNSON, H.E., JR. CORP. RES. AND DEV., GEN. ELECTR. CO., SCHENECTADY, NY, USA HUM. SYST. MANAGE. (NETHERLANDS) VOL.4, NO.4 255-62 AUTUMN 1984 CODEN: HSMADU ISSN: 0167-2533 U. S. Copyright Clearance Center Code: 0167-2533/84/$3.00 Treatment: PRACTICAL Doc. Type: JRNL PAP. Lang.: ENG. (15 Refs) GENERAL ELECTRIC COMPANY'S CORPORATE RESEARCH AND DEVELOPMENT HAS APPLIED EXPERT SYSTEM TECHNOLOGY TO THE PROBLEM OF TROUBLESHOOTING AND THE REPAIR OF DIESEL ELECTRIC LOCOMOTIVES IN RAILROAD 'RUNNING REPAIR SHOPS'. THE EXPERT SYSTEM USES PRODUCTION RULES AND AN INFERENCE ENGINE THAT CAN DIAGNOSE MULTIPLE PROBLEMS WITH THE LOCOMOTIVE AND CAN SUGGEST REPAIR PROCEDURES TO MAINTENANCE PERSONNEL. A PROTOTYPE SYSTEM HAS BEEN IMPLEMENTED IN FORTH, RUNNING ON A DIGITAL EQUIPMENT PDP 11/23 UNDER RSX-11M. THIS SYSTEM CONTAINS APPROXIMATELY 530 RULES (ROUGHLY 330 RULES FOR THE TROUBLESHOOTING SYSTEM, AND 200 RULES FOR THE HELP SYSTEM), PARTIALLY REPRESENTING THE KNOWLEDGE OF A SENIOR FIELD SERVICE ENGINEER. THE INFERENCE ENGINE USES A MIXED-MODE CONFIGURATION, CAPABLE OF RUNNING IN EITHER THE FORWARD OR BACKWARD MODE. THE HELP SYSTEM CAN PROVIDE THE OPERATOR WITH ASSISTANCE BY DISPLAYING TEXTUAL INFORMATION, CAD DIAGRAMS OR REPAIR SEQUENCES FROM A VIDEO DISK. THE RULES ARE WRITTEN IN A REPRESENTATION LANGUAGE CONSISTING OF NINE PREDICATE FUNCTIONS, EIGHT VERBS, AND FIVE UTILITY FUNCTIONS. THE FIRST FIELD PROTOTYPE EXPERT SYSTEM, DESIGNATED CATS-1 (COMPUTER-AIDED TROUBLESHOOTING SYSTEM-VERSION 1), WAS DELIVERED IN JULY 1983 AND IS CURRENTLY UNDER FIELD EVALUATION. Desc.: EXPERT SYSTEMS; MAINTENANCE ENGINEERING; DIESEL-ELECTRIC LOCOMOTIVES Ident.: DIESEL ELECTRIC LOCOMOTIVE; EXPERT SYSTEM; TROUBLESHOOTING; REPAIR; RUNNING REPAIR SHOPS; PRODUCTION RULES; INFERENCE ENGINE; MAINTENANCE PERSONNEL; FORTH; DIGITAL EQUIPMENT PDP 11/23; HELP SYSTEM; CATS-1; COMPUTER-AIDED TROUBLESHOOTING SYSTEM Class Codes: C7440; C7410 Y034013 110 1379964 C85005582 OFF-LINE PLANNING OF COLLISION-FREE TRAJECTORIES AND OBJECT GRASPING FOR MANIPULATION ROBOTS JENTSCH, W. CENTRAL INST. FOR CYBERN. AND INF. PROCESSES, AKAD. DER WISSENSCHAFTEN, BERLIN, GERMANY PLANDER, I. (Editors) Sponsor: ACAD. SCI. USSR; SLOVAK ACAD. SCI ARTIFICIAL INTELLIGENCE AND INFORMATION-CONTROL SYSTEMS OF ROBOTS. PROCEEDINGS OF THE THIRD INTERNATIONAL CONFERENCE 193-6 1984 11-15 JUNE 1984 SMOLENICE, CZECHOSLOVAKIA Publ: NORTH-HOLLAND, AMSTERDAM, NETHERLANDS XVI+413 pp. ISBN 0 444 87533 6 Treatment: THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (7 Refs) THE PROBLEM OF GEOMETRICAL PLANNING FOR MANIPULATION ROBOTS IS CONSIDERED. A DESCRIPTION OF THE OBSTACLES IN THE CONFIGURATION SPACE IS OBTAINED BY A SELF-LEARNING CLASSIFICATION METHOD. THE PATH-FIND PROBLEM IS RESOLVED BY A SPECIAL PATH-GRAPH CONSTRUCTION AND AN A*-SEARCH. THE GENERATION OF SUITABLE GRASPING AND REGRASPING VECTORS IS BASED ON THE CORRESPONDENCE BETWEEN CONFIGURATION AND WORK SPACE. Desc.: INDUSTRIAL ROBOTS Ident.: COLLISION-FREE TRAJECTORIES; OBJECT GRASPING; MANIPULATION ROBOTS; GEOMETRICAL PLANNING; OBSTACLES; SELF-LEARNING CLASSIFICATION; PATH-FIND PROBLEM Class Codes: C3355; C1230 Y034013 111 1379294 C85004664 A PLAN GENERATING SYSTEM OF AN INTELLECTUAL ROBOT USING A FRAME REPRESENTATION OF THE INFORMATION ZLATAREVA, N.P. INST. OF IND. CYBERN. AND ROBOTICS, BULGARIAN ACAD. OF SCI., SOFIA, BULGARIA PLANDER, I. (Editors) Sponsor: ACAD. SCI. USSR; SLOVAK ACAD. SCI ARTIFICIAL INTELLIGENCE AND INFORMATION-CONTROL SYSTEMS OF ROBOTS. PROCEEDINGS OF THE THIRD INTERNATIONAL CONFERENCE 399-403 1984 11-15 JUNE 1984 SMOLENICE, CZECHOSLOVAKIA Publ: NORTH-HOLLAND, AMSTERDAM, NETHERLANDS XVI+413 pp. ISBN 0 444 87533 6 Treatment: THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (2 Refs) TRADITIONALLY, THE TASK OF THE PLAN GENERATING SYSTEM IS DEFINED AS THE TASK OF HEURISTIC SEARCH, WHERE THE INITIAL MODEL OF THE ENVIRONMENT IS FORMULATED, FOLLOWED BY THE FINAL MODEL IN WHICH THE DESIRED AIM IS CONSIDERED TO BE REACHED AND A SET OF OPERATORS WHICH TRANSFORM ONE ENVIRONMENTAL MODEL INTO ANOTHER. FORMALLY, THIS TASK MAY BE REPRESENTED AS A FOURTUPLE. Desc.: ROBOTS Ident.: PLAN GENERATING SYSTEM; INTELLECTUAL ROBOT; FRAME REPRESENTATION; HEURISTIC SEARCH Class Codes: C1230 Y034013 112 1379255 C85004625 APPROXIMATE REASONING AS A BASIS FOR RULE-BASED EXPERT SYSTEMS YAGER, R.R. MACHINE INTELLIGENCE INST., IONA COLL., NEW ROCHELLE, NY, USA IEEE TRANS. SYST., MAN AND CYBERN. (USA) VOL.SMC-14, NO.4 636-43 JULY-AUG. 1984 CODEN: ISYMAW ISSN: 0018-9472 U. S. Copyright Clearance Center Code: 0018-9472/84/0700-0636$01.00 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (18 Refs) THE CONCEPT OF A RULE-BASED EXPERT SYSTEM THAT INCLUDES DATA AND PRODUCTION RULES IS DISCUSSED. IT IS SHOWN HOW THE THEORY OF APPROXIMATE REASONING DEVELOPED BY L.A. ZADEH (NEW YORK, WILEY, 1979) PROVIDES A NATURAL FORMAT FOR REPRESENTING THE KNOWLEDGE AND PERFORMING THE INFERENCES IN THE RULE-BASED EXPERT SYSTEMS. THE REPRESENTATION ABILITY OF THE SYSTEMS IS EXTENDED BY PROVIDING A NEW STRUCTURE FOR INCLUDING THE RULES THAT ONLY REQUIRE THE SATISFACTION TO SOME SUBSET OF THE REQUIREMENTS IN ITS ANTECEDENT. THIS IS ACCOMPLISHED BY USE OF FUZZY QUANTIFIERS. A METHODOLOGY IS ALSO PROVIDED FOR THE INCLUSION OF A FORM OF UNCERTAINTY IN THE EXPERT SYSTEM ASSOCIATED WITH THE BELIEF ATTRIBUTED TO THE DATA AND PRODUCTION RULES. Desc.: EXPERT SYSTEMS; APPROXIMATION THEORY; FUZZY SET THEORY Ident.: RULE-BASED EXPERT SYSTEMS; EXPERT SYSTEM; APPROXIMATE REASONING; KNOWLEDGE; INFERENCES; SUBSET; FUZZY QUANTIFIERS; UNCERTAINTY; PRODUCTION RULES Class Codes: C1230; C4130; C1160; C7000 Y034013 113 1379251 C85004621 SEARCH FOR SOLUTIONS USING KNOWLEDGE SUDEYKIN, M.I. TEKH. KIBERN. (USSR) VOL.21, NO.4 100-11 JULY-AUG. 1983 CODEN: TEKIB8 Trans in: ENG. CYBERN. (USA) VOL.21, NO.4 1-11 JULY-AUG. 1983 CODEN: ENCYAF ISSN: 0013-788X U. S. Copyright Clearance Center Code: 0013-788X/83/0004-0001$07.50/0 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (6 Refs) THE SEARCH FOR SOLUTIONS IN THE INTEGRATED ROBOT SYSTEM 'SPRINT' FOR PLANNING OF OPERATIONS IS REALIZED USING STORAGE IN THE FUNCTIONING OF THE KNOWLEDGE SYSTEM. THE BASIS OF KNOWLEDGE ORGANIZATION IN SPRINT IS A TWO-LEVEL GRAMMAR OF SPECIAL FORM. TRAINING IS CARRIED OUT IN SPRINT BY CHANGING THE STRUCTURE OF THE OBTAINED SOLUTION AND BY GENERALIZING THE STRUCTURES. THE BASIS OF THE STRUCTURAL TRANSFORMATIONS IS A REDUCTION GRAMMAR INTENDED FOR DESCRIBING THE PROCESS OF TRANSFORMING SYNTACTIC STRUCTURES. THE SEARCH FOR SOLUTIONS IN SPRINT IS CARRIED OUT BY REDUCING PROBLEMS TO SUBPROBLEMS, THE CONTROL OF WHICH IS BASED ON THE USE OF KNOWLEDGE. Desc.: ROBOTS; KNOWLEDGE ENGINEERING; LEARNING SYSTEMS; GRAMMARS Ident.: SOLUTION SEARCHING; ROBOT SYSTEM; KNOWLEDGE SYSTEM; KNOWLEDGE ORGANIZATION; SPRINT; TWO-LEVEL GRAMMAR; STRUCTURAL TRANSFORMATIONS; REDUCTION GRAMMAR; SYNTACTIC STRUCTURES Class Codes: C1230; C4210 Y034013 114 1379249 C85004619 THE APPLICATION OF TECHNIQUES OF ARTIFICIAL INTELLIGENCE TO TIMETABLE SCHEDULING LIN ZHANGXI; LIN YAORUI DEPT. OF COMPUT. ENG. AND SCI., TSINGHUA UNIV., CHINA J. TSINGHUA UNIV. (CHINA) VOL.24, NO.2 1-10 1984 CODEN: CHHPAH ISSN: 0577-9189 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: CHINESE (5 Refs) A TIMETABLE SCHEDULING SYSTEM TISER, WHICH IS PROGRAMMED BY LISP AND BASED ON THE PRINCIPLE OF BIMULTIGRAPH MATCHING, IS PRESENTED. USING THE ARTIFICIAL INTELLIGENCE TECHNIQUES TISER HEURISTICALLY SEARCHES THE AND/OR GRAPHS PRODUCED DURING SCHEDULING AND SUCCESSFULLY WORKS OUT SOME TIMETABLES FROM TWO GROUPS OF TESTING DATA. Desc.: ARTIFICIAL INTELLIGENCE; SCHEDULING; HEURISTIC PROGRAMMING Ident.: ARTIFICIAL INTELLIGENCE; TIMETABLE SCHEDULING; TISER; PROGRAMMED BY LISP; BIMULTIGRAPH MATCHING; AND/OR GRAPHS Class Codes: C1230; C7100 Y034013 115 1379247 C85004617 A MODIFIED UNIFICATION PROCEDURE MOROKHOVETS, M.K. KIBERNETIKA (USSR) VOL.20, NO.1 96-9 JAN.-FEB. 1984 CODEN: KBRNA5 ISSN: 0023-1274 Trans in: CYBERNETICS (USA) VOL.20, NO.1 147-52 JAN.-FEB. 1984 CODEN: CYBNAW ISSN: 0011-4235 U. S. Copyright Clearance Center Code: 0011-4235/84/2001-0147$08.50 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (2 Refs) CONSIDERS A METHOD OF APPLICATION OF A NATURAL MATHEMATICAL TECHNIQUE TO PROCEDURES OF DERIVATION SEARCH-SPECIFICALLY, THE USE OF CONNECTIONS OF MATHEMATICAL CONCEPTS IN PROOFS. THIS TECHNIQUE CAN BE FORMALIZED BY USING A SPECIAL LANGUAGE OF REPRESENTATION OF MATHEMATICAL STATEMENTS, WHICH IS USED AS THE DATA LANGUAGE IN SAD SYSTEMS. THIS LANGUAGE HAS ANALOGS OF CONCEPTS CALLED CONSTRUCTS AND IS OBTAINED BY SUBSTITUTING STATEMENTS INTO THE SO-CALLED CONSTRUCT FORMANTS. Desc.: SET THEORY; THEOREM PROVING Ident.: SET THEORY; THEOREM PROVING; MODIFIED UNIFICATION PROCEDURE DERIVATION SEARCH; CONNECTIONS OF MATHEMATICAL CONCEPTS; CONSTRUCTS Class Codes: C1230; C1160 Y034013 116 1364744 C85003890 AN EXPERT SYSTEM SCENARIO FOR COMPUTER-AIDED CONTROL ENGINEERING TAYLOR, J.H.; FREDERICK, D.K.; JAMES, J.R. GENERAL ELECTRIC CO., SCHENECTADY, NY, USA Sponsor: AMERICAN AUTOM. CONTROL COUNCIL PROCEEDINGS OF THE 1984 AMERICAN CONTROL CONFERENCE (IEEE CAT. NO. 84CH2024-8) 120-8 VOL.1 1984 6-8 JUNE 1984 SAN DIEGO, CA, USA Publ: IEEE, NEW YORK, USA 3 VOL. 1913 pp. Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (14 Refs) AN EXPERT SYSTEM TO CREATE THIRD-GENERATION MAN/MACHINE ENVIRONMENTS FOR COMPUTER-AIDED CONTROL ENGINEERING IS PROPOSED. THIS APPROACH PROMISES TO PROVIDE A HIGH-LEVEL DESIGN ENVIRONMENT THAT IS POWERFUL, SUPPORTIVE, FLEXIBLE, BROAD IN SCOPE, AND READILY ACCESSIBLE TO NONEXPERT USERS. THE CAPABILITIES OF SUCH AN EXPERT SYSTEM VIA AN EXTENDED TRANSACTION (SAMPLE INTERACTION BETWEEN THE EXPERT SYSTEM AND A USER) ARE ILLUSTRATED. Desc.: ARTIFICIAL INTELLIGENCE; CONTROL ENGINEERING COMPUTING; MAN-MACHINE SYSTEMS Ident.: ARTIFICIAL INTELLIGENCE; EXPERT SYSTEM; COMPUTER-AIDED CONTROL ENGINEERING; THIRD-GENERATION MAN/MACHINE ENVIRONMENTS Class Codes: C7420 Y034013 117 1364452 C85003283 THE DEVELOPMENT OF AN EXPERT SYSTEM TO ASSESS THE VULNERABILITY OF ASSETS TO SUBNATIONAL ADVERSARIES HUNT, A.R.; MULLEN, R.K. GENERAL RES. CORP., SANTA BARBARA, CA, USA JACKSON, J.S.; DE VORE, R.W. (Editors) Sponsor: UNIV. KENTUCKY; IEEE PROCEEDINGS OF THE 1984 CARNAHAN CONFERENCE ON SECURITY TECHNOLOGY (CAT. NO. 84CH2023-0) 203-8 1984 16-18 MAY 1984 LEXINGTON, KY, USA Publ: IEEE, NEW YORK, USA 217 pp. ISBN 0 89779 059 6 Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (8 Refs) DESCRIBES THE DEVELOPMENT OF THE INTELLIGENT MACHINE MODEL COMPUTER PROGRAM (TIMM), WHICH OVERCOMES THE COMPUTER/EXPERT INTERFACE PROBLEMS OF CONVENTIONAL SYSTEM. TIMM IS A DOMAIN-INDEPENDENT, KNOWLEDGE-BASED PROGRAM WHICH PERMITS A USER TO CONSTRUCT, EXERCISE, MODIFY, AND INTERROGATE A LARGE BODY OF KNOWLEDGE WHOSE SUBJECT MATTER IS SELECTED BY THE USER. TIMM LEADS THE HUMAN EXPERT TO DESCRIBE THE PROBLEM SPACE IN TERMS OF ITS ANSWERING QUESTIONS ABOUT SPECIFIC COMBINATIONS OF FACTOR VALUES. THIS TRAINING REQUIRES NO COMPUTER KNOWLEDGE FROM THE EXPERT. THE SYSTEM AUTOMATICALLY GENERATES PLAUSIBLE EVENTS IN THE PROBLEM DOMAIN, QUERYING THE EXPERT AS TO THE CORRECT DECISION IN THE PRESENTED CASE. THUS, THE EXPERT USES HIS OR HER DECISION-MAKING EXPERTISE DIRECTLY. RESULTS HAVE SHOWN THAT TIMM DOES MAKE IT MUCH EASIER FOR AN EXPERT TO SIT IN FRONT OF A COMPUTER AND CODIFY HIS KNOWLEDGE EFFICIENTLY. THE PROTOTYPE EXPERT SYSTEM ACHIEVED A PERFORMANCE WHICH AGREED WITH THE JUDGEMENT OF A HUMAN EXPERT NEARLY 60PERCENT OF THE TIME AFTER A TRAINING PERIOD OF ONLY TWO WEEKS. Desc.: ARTIFICIAL INTELLIGENCE; MILITARY SYSTEMS; MILITARY COMPUTING Ident.: KNOWLEDGE-BASED SYSTEM; EXPERT SYSTEM; INTELLIGENT MACHINE MODEL COMPUTER PROGRAM; TIMM; TRAINING Class Codes: C7150; C1230 Y034013 118 1350882 B84060127, C84048345 'IGNORANCE-BASED' SYSTEMS GEVINS, A.S.; MORGAN, N.H. ADAPTIVE DECISION INSTRUMENTS, SAN FRANCISCO, CA, USA Sponsor: IEEE ICASSP 84. PROCEEDINGS OF THE IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING 39A.5/1-4 VOL.3 1984 19-21 MARCH 1984 SAN DIEGO, CA, USA Publ: IEEE, NEW YORK, USA 3 VOL. (742+864+664) pp. U. S. Copyright Clearance Center Code: CH1945-5/84/0000-0424$01.00 Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (35 Refs) RECENT WORK IN ARTIFICIAL INTELLIGENCE AND SIGNAL PROCESSING HAS SUGGESTED THAT A MERGER OF THE TWO FIELDS WOULD BE PROFITABLE FOR BOTH. THE AUTHORS DISCUSS THE IMPROVEMENT OF DECISION SYSTEMS BY THE USE OF ITERATIVE MATHEMATICAL TECHNIQUES. THESE TECHNIQUES COULD BE CALLED 'IGNORANCE-BASED' SINCE THEY CAN BE CHARACTERIZED BY EXHAUSTIVE SEARCHES FOR USEFUL COMBINATIONS OF PROBLEM-RELEVANT VARIABLES IN DATA SPACES FOR WHICH HUMAN KNOWLEDGE IS INCOMPLETE. Desc.: ARTIFICIAL INTELLIGENCE; COMPUTERISED SIGNAL PROCESSING; COMPUTERISED PICTURE PROCESSING; SPEECH ANALYSIS AND PROCESSING Ident.: IGNORANCE BASED SYSTEMS; PATTERN RECOGNITION; SPEECH CLASSIFICATION; ARTIFICIAL INTELLIGENCE; SIGNAL PROCESSING; DECISION SYSTEMS; ITERATIVE MATHEMATICAL TECHNIQUES Class Codes: B6140; B6130; C1260; C1250; C1230 Y034013 119 1345950 C84048114 APPLICATIONS OF ARTIFICIAL INTELLIGENCE IN AUTOMATED ROUTE PLANNING DENTON, R.V.; FROEBERG, P.L. TAU CORP., LOS GATOS, CA, USA Sponsor: SPIE PROC. SPIE INT. SOC. OPT. ENG. (USA) VOL.485 126-32 1984 CODEN: PSISDG ISSN: 0277-786X APPLICATIONS OF ARTIFICIAL INTELLIGENCE 3-4 MAY 1984 ARLINGTON, VA, USA Treatment: THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (3 Refs) ADDRESSES THE PROBLEM OF ROUTE PLANNING FOR GROUND VEHICLES. THE RELEVANCE OF EXPERT SYSTEMS AND OTHER TECHNIQUES FOR ROUTE PLANNING IS DISCUSSED. A PARTICULARLY EFFICIENT SEARCH PROCEDURE IS APPLIED TO SEVERAL EXAMPLE COMPOSITE MAPS TO DEMONSTRATE THE POWER OF THE APPROACH. Desc.: ARTIFICIAL INTELLIGENCE; CARTOGRAPHY; ROAD VEHICLES Ident.: ARTIFICIAL INTELLIGENCE; AUTOMATED ROUTE PLANNING; GROUND VEHICLES; EXPERT SYSTEMS; SEARCH PROCEDURE; COMPOSITE MAPS Class Codes: C1230; C3360B; C1290H Y034013 120 1335905 A84093951, B84056139 RULE-BASED METHODS FOR ELECTROENCEPHALOGRAM EVALUATION BOURNE, J.R. DEPT. OF ELECTRICAL AND BIOMEDICAL ENGNG., VANDERBILT UNIV., NASHVILLE, TN, USA GERHARD, G.C.; MILLER, W.T. (Editors) Sponsor: IEEE FRONTIERS OF ENGINEERING AND COMPUTING IN HEALTH CARE - 1983. PROCEEDINGS OF THE FIFTH ANNUAL CONFERENCE 717 1983 10-12 SEPT. 1983 COLUMBUS, OH, USA Publ: IEEE, NEW YORK, USA 735 pp. Treatment: GENERAL, REVIEW Doc. Type: CONFERENCE PAPER Lang.: ENG. SUMMARY FORM ONLY GIVEN, AS FOLLOWS. A DESCRIPTION IS PROVIDED OF FUNDAMENTAL CONCEPTS OF RULE-BASED SYSTEMS TECHNOLOGY AND PRACTICAL METHODS FOR APPLYING THESE CONCEPTS TO ELECTROENCEPHALOGRAM (EEG) INTERPRETATION AND ALLIED PROBLEMS. FIRST, THE USE OF RULES IN SYNTACTIC PATTERN RECOGNITION FOR DATA-DRIVEN ASSESSMENT OF PATTERN CHARACTERISTICS IS DESCRIBED, INCLUDING SEVERAL EXAMPLES. METHODS FOR IDENTIFICATION OF FUNDAMENTAL SIGNAL CHARACTERISTICS AND PARSING STRATEGIES FOR WAVEFORM EVALUATION ARE OUTLINED AS WELL AS A SUMMARY OF THE PERCEIVED UTILITY OF THESE TECHNIQUES FOR EEG ASSESSMENT. NEXT, RULE SYSTEMS FOR ENCODING SEMANTIC KNOWLEDGE ARE EXAMINED AND AN EXAMPLE OF A BACKWARD-CHAINING GOAL-DRIVEN SYSTEM FOR MULTICHANNEL EEG EVALUATION IS REPORTED. THE DESIGN OF AN EXPERT SYSTEM FOR SIMULATION OF CLINICAL KNOWLEDGE, INCLUDING EEG KNOWLEDGE, IS PRESENTED. A CRITICAL REVIEW OF PROBLEMS IN THE DESIGN OF SUCH SYSTEM IS GIVEN NEXT. DESIGN ISSUES CONCERNED WITH CONTROL STRATEGIES, KNOWLEDGE ACQUISITION, REPRESENTATION, AND COMPILATION ARE EXAMINED. THE ADVANTAGES OF EXPLANATION MODULES AND NATURAL LANGUAGE USER INTERFACES ARE DISCUSSED. PROGRAMMING METHODS ARE EXPLORED, AND A PROSPECTUS FOR RESEARCH USING THE RULE-BASED SYSTEMS TECHNOLOGY IS PRESENTED. CURRENT PROBLEMS ARE HIGHLIGHTED AND SUGGESTIONS FOR FRUITFUL RESEARCH DIRECTIONS GIVEN. Desc.: ELECTROENCEPHALOGRAPHY Ident.: SEMANTIC KNOWLEDGE ENCODING; SYNTACTIC PATTERN RECOGNITION; FUNDAMENTAL SIGNAL CHARACTERISTICS; WAVEFORM EVALUATION; BACKWARD-CHAINING GOAL-DRIVEN SYSTEM; CONTROL STRATEGIES; KNOWLEDGE ACQUISITION; REPRESENTATION; COMPILATION; EXPLANATION MODULES; NATURAL LANGUAGE USER INTERFACES Class Codes: A8770; A8730C; B7510D Y034013 121 1330880 B84050681, C84041907 ON THE QUALITY OF HEURISTIC SOLUTIONS TO A 19*19 QUADRATIC ASSIGNMENT PROBLEM BRUIJS, P.A. DR. NEHER LABS., PTT, LEIDSCHENDAM, NETHERLANDS EUR. J. OPER. RES. (NETHERLANDS) VOL.17, NO.1 21-30 JULY 1984 CODEN: EJORDT ISSN: 0377-2217 U. S. Copyright Clearance Center Code: 0377-2217/84/$3.00 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (17 Refs) RECENTLY MURTAGH ET AL. COMMUNICATED THE BEST (?) KNOWN SOLUTION TO A 19*19 QUADRATIC ASSIGNMENT PROBLEM, PRESENTED AND SOLVED EARLIER BY ELSHAFEI. THEIR APPROACHES ARE BASED ON THE APPLICATION OF QUADRATIC PROGRAMMING AND ON THE CONSTRUCTION OF A GOOD INITIAL SOLUTION RESPECTIVELY. THEIR SOLUTIONS ARE CONFRONTED WITH ONE BETTER SOLUTION OBTAINED BY LASHKARI AND JAISINGH BY MEANS OF LINEAR ASSIGNMENT APPROXIMATIONS, AND MANY BETTER SOLUTIONS OBTAINED BY THE PRESENT AUTHOR THROUGH THE APPLICATION OF EXCHANGE HEURISTICS. THE RESULTS STRESS THE NONCONVEX CHARACTER OF THE PROBLEM, AND SHOW THE RATHER MISLEADING NATURE OF THE CONCEPT: 'GOOD INITIAL SOLUTION'. THE QUALITY OF INITIAL AND FINAL SOLUTION PROVE TO BE RATHER WEAKLY (OR EVEN NEGATIVELY!) CORRELATED. A SIMPLE, BUT BROAD AND RANDOMIZED SEARCH RESULTS IN SIGNIFICANTLY BETTER SOLUTIONS, WHILE SIMPLE STATISTICAL AIDS PROVIDE AN ESTIMATE OF THE QUALITY OF THESE SOLUTIONS. FINALLY, THE AUTHOR MAKES REMARKS CONCERNING SOME PRACTICAL ISSUES. Desc.: QUADRATIC PROGRAMMING; HEURISTIC PROGRAMMING Ident.: QUALITY OF HEURISTIC SOLUTIONS; 19*19 QUADRATIC ASSIGNMENT PROBLEM; QUADRATIC PROGRAMMING; GOOD INITIAL SOLUTION; EXCHANGE HEURISTICS Class Codes: B0260; C1180; C1230 Y034013 122 1330179 C84047250 THE FIRST CAD/CAM EXPERT SYSTEMS RUEHER, M. ELECTRON. IND. (FRANCE) NO.70 55-8 1 MAY 1984 CODEN: EIDUDA ISSN: 0244-903X Treatment: PRACTICAL Doc. Type: JRNL PAP. Lang.: FRENCH (9 Refs) EXAMPLES ARE GIVEN OF KNOWLEDGE-BASED SYSTEMS FOR DESIGN OF MECHANICAL AND ELECTROMECHANICAL PRODUCTS, MACHINING APPARATUS, AND ELECTRONIC CIRCUITS, AND TWO EXPERT SYSTEMS FOR PRODUCING MANUFACTURING PLANS (ONE FOR AUTOMATING MACHINING RANGES, AND THE OTHER FOR CONTROL OF A MOVING ROBOT). THEIR MAIN WEAKNESSES ARE FOUND AT THE LEVELS OF DATA MANAGEMENT AND USER DIALOGUE. Desc.: CAD; MANUFACTURING COMPUTER CONTROL; ARTIFICIAL INTELLIGENCE Ident.: COMPUTER-AIDED DESIGN; ARTIFICIAL INTELLIGENCE; COMPUTER-AIDED MANUFACTURING; CAD; CAM; EXPERT SYSTEMS; KNOWLEDGE-BASED SYSTEMS; DATA MANAGEMENT; USER DIALOGUE Class Codes: C7440; C3350E; C7420; C7410D Y034013 123 1330046 C84046882 ROTES: AN EXPERT SYSTEM FOR ROBOT TROUBLESHOOTING KAWAMURA, K.; CLOSE, J.J. SCHOOL OF ENGNG., VANDERBILT UNIV., NASHVILLE, TN, USA Sponsor: IEEE; MISSISSIPPI STATE UNIV PROCEEDINGS OF THE SIXTEENTH SOUTHEASTERN SYMPOSIUM ON SYSTEM THEORY 129-32 1984 25-27 MARCH 1984 MISSISSIPPI STATE, MS, USA Publ: IEEE COMPUT. SOC. PRESS, SILVER SPRING, MD, USA VII+237 pp. ISBN 0 8186 0556 1 U. S. Copyright Clearance Center Code: 0094-2898/84/0000-0129$01.00 Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (6 Refs) THE APPLICATION OF EXPERT SYSTEM TECHNOLOGY TO THE TROUBLESHOOTING AND MAINTENANCE OF THE CINCINNATI-MILACRON T3 ROBOT IS REPORTED. ROTES USES PRODUCTION RULES, AN INFERENCE ENGINE CALLED GENIE, AND A TREE-STRUCTURED HELP SYSTEM CALLED GENHELS. GENHELS ACCEPTS USER INPUT IN THE FORM OF A DATABASE SEARCH KEY AND THE OUTPUT'S TEXTURAL INFORMATION ASSOCIATED WITH THAT KEY. A PROTOTYPE EXPERT SYSTEM IS BEING IMPLEMENTED ON THE DEC 1099 SYSTEM. Desc.: ARTIFICIAL INTELLIGENCE; MAINTENANCE ENGINEERING Ident.: ARTIFICIAL INTELLIGENCE; ROTES; EXPERT SYSTEM; ROBOT TROUBLESHOOTING; CINCINNATI-MILACRON T3 ROBOT; PRODUCTION RULES; INFERENCE ENGINE; GENIE; TREE-STRUCTURED HELP SYSTEM; GENHELS; DATABASE SEARCH KEY; DEC 1099 SYSTEM Class Codes: C7400; C1230 Y034013 124 1329453 C84046084 RESEARCH DIRECTIONS IN INFORMATION SYSTEMS KONSYNSKI, B.R. MANAGEMENT INFORMATION SYSTEMS DEPT., UNIV. OF ARIZONA, TUCSON, AZ, USA CANADIAN INFORMATION PROCESSING SOCIETY SESSION 84 PROCEEDINGS. 1984: IMAGES OF FEAR, IMAGES OF HOPE 252-8 1984 9-11 MAY 1984 CALGARY, ALTA., CANADA Publ: CANADIAN INF. PROCESSING SOC., TORONTO, ONT., CANADA XVI+513 pp. Treatment: PRACTICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (7 Refs) THE PAPER OFFERS A FRAMEWORK FOR OBSERVING SEVERAL MAJOR TRENDS IN SEVERAL RESEARCH DIRECTIONS IN THE FIELD OF INFORMATION SYSTEMS. SEVERAL FREQUENTLY OCCURRING PARADIGMS ARE REVIEWED AND FUTURE RESEARCH TRENDS ARE HIGHLIGHTED. TRENDS IN DATA MANAGEMENT, EXPERT SYSTEMS, DECISION SUPPORT, SYSTEMS DEVELOPMENT, STRATEGIC USE OF SYSTEMS, USER INTERFACE DESIGN AND OTHER DIMENSIONS OF CHANGE ARE DISCUSSED. FOUNDATION CONCEPTS ARE DISCUSSED . Desc.: MANAGEMENT INFORMATION SYSTEMS; SYSTEMS ANALYSIS Ident.: INFORMATION SYSTEMS; DATA MANAGEMENT; EXPERT SYSTEMS; DECISION SUPPORT; SYSTEMS DEVELOPMENT; USER INTERFACE DESIGN Class Codes: C7100; C6110 Y034013 125 1326905 C84042047 NON-CANONICAL SIMPLIFICATION KODRATOFF, Y.; COSTA, E. UNIV. PARIS-SUD, ORSAY, FRANCE TRAPPL, R. (Editors) CYBERNETICS AND SYSTEMS RESEARCH 2. PROCEEDINGS OF THE SEVENTH EUROPEAN MEETING 705-13 1984 24-27 APRIL 1984 VIENNA, AUSTRIA Publ: NORTH-HOLLAND, AMSTERDAM, NETHERLANDS XIV+840 pp. ISBN 0 444 86898 4 Treatment: THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (18 Refs) MANY ARTIFICIAL INTELLIGENCE EFFORTS ARE STOPPED BECAUSE SOME INVENTIVE STEP IS NEEDED AT A POINT OF THE 'INTELLIGENT' PROCESS UNDER SIMULATION. THIS INVENTIVE STEP MAY BE REALIZED BY AN EXHAUSTIVE SEARCH OF ALL THE POSSIBILITIES IN ORDER TO ATTEMPT ALL THE POSSIBLE GUESSES SOME OF WHICH LEADING TO A SOLUTION. IN SOME WELL-DEFINED CASES KNUTH-BENDIX COMPLETION IS A PERFECT TOOL IN ORDER TO ACHIEVE THIS GOAL. THE AUTHORS PRESENT A METHODOLOGY WHICH IS STILL APPLICABLE EVEN WHEN THE PROBLEM IS BADLY DEFINED. THE METHODOLOGY PRESENTED DOES NOT AVOID ALL THE COMBINATORIAL SEARCH BUT REDUCES ITS EXTENT. Desc.: ARTIFICIAL INTELLIGENCE; THEOREM PROVING Ident.: NONCANONICAL SIMPLIFICATION; ARTIFICIAL INTELLIGENCE; KNUTH BENDIX COMPLETION; METHODOLOGY; COMBINATORIAL SEARCH Class Codes: C1230 Y034013 126 1314353 A84089603, B84047812, C84037554 SIMULATION OF PHONEMIC ERRORS USING ARTIFICIAL INTELLIGENCE SYMBOL PROCESSING TECHNIQUES REGGIA, J.A.; AHUJA, S.B. UNIV. OF MARYLAND, COLLEGE PARK, MD, USA COMER, E.R. (Editors) Sponsor: IEEE RECORD OF PROCEEDINGS OF THE 17TH ANNUAL SIMULATION SYMPOSIUM 49-64 1984 14-16 MARCH 1984 TAMPA, FL, USA Publ: IEEE, NEW YORK, USA VII+267 pp. ISBN 0 8186 0527 8 Treatment: THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (12 Refs) USING HEURISTICALLY GUIDED STATE-SPACE SEARCH, A PROGRAM HAS BEEN DEVELOPED TO SIMULATE PHONEMIC ERRORS OCCURRING IN THE SPEECH OF BOTH NORMAL INDIVIDUALS AND NEUROLOGICALLY IMPAIRED PATIENTS. SIMULATIONS ARE BASED ON AN INTERCHANGEABLE RULE/OPERATOR SET OF ERRORS WHICH REPRESENT A THEORY OF PHONEMIC PROCESSING FAULTS. Desc.: SPEECH ANALYSIS AND PROCESSING; ARTIFICIAL INTELLIGENCE; DIGITAL SIMULATION; MEDICAL COMPUTING Ident.: PHONEMIC ERRORS; ARTIFICIAL INTELLIGENCE; SYMBOL PROCESSING STATE-SPACE SEARCH; SPEECH; PHONEMIC PROCESSING FAULTS Class Codes: A8736; A4370; B6130; C1230; C7330; C5585 Y034013 127 1312206 C84041038 CONTROL AT THE CROSSROADS SHAW, R. IND. MANAGE. + DATA SYST. (GB) 10-12 MARCH-APRIL 1984 CODEN: IMDSD8 ISSN: 0007-6929 Treatment: GENERAL, REVIEW Doc. Type: JRNL PAP. Lang.: ENG. JUST NOW, ELECTRONIC CONTROL TECHNOLOGY IS AT SOMETHING OF A CROSSROADS. IT ARISES FROM THE VERY PROFUSION OF AVAILABLE TECHNOLOGY AND THE MIND-BOGGLING COMPLEXITY OF SYSTEMS WHICH CAN RESULT. THERE IS NO DOUBT THAT A MODEST AMOUNT OF AUTOMATION BRINGS ECONOMIC BENEFIT TO THE USER. BUT IT IS NOW LIKELY THAT AUTOMATION JUST ONE OR TWO STEPS BEYOND PRESENT PRACTICE WILL BEGIN TO REQUIRE THE USE OF EXPERT SYSTEMS TO MAKE SYSTEM OPERATION INTELLIGIBLE TO THE ORDINARY HUMAN OPERATOR. THE DILEMMA FOR THE USER IS SHOULD HE WAIT FOR NEXT YEAR'S MODEL? THE DILEMMA FOR THE SUPPLIER IS WHICH PRODUCT SHOULD HE DESIGN AND MARKET?. Desc.: COMPUTERISED CONTROL; CONTROL ENGINEERING COMPUTING Ident.: ELECTRONIC CONTROL TECHNOLOGY; AUTOMATION; ECONOMIC BENEFIT EXPERT SYSTEMS; SYSTEM OPERATION Class Codes: C7420; C3000 Y034013 128 1311076 C84039534 SOFTWARE'S NEXT DIMENSION HORTON, F.W., JR. COMPUTERWORLD (USA) VOL.18, NO.26 ID/7-11 25 JUNE 1984 CODEN: CMPWAB ISSN: 0010-4841 Treatment: GENERAL, REVIEW Doc. Type: JRNL PAP. Lang.: ENG. PRODUCT DEVELOPERS OF ARTIFICIAL INTELLIGENCE, EXPERT SYSTEMS AND GATEWAY SOFTWARE (FOR SEARCHING COMMERCIAL DATABASES) ARE NOT THINKING BOLDLY ENOUGH. SOFTWARE EXPLOITS THE POWER AND VERSATILITY OF MACHINES. IT NOT ONLY HELPS US 'DO THE THING RIGHT', IT ALSO HELPS US 'DO THE THING WELL'. BUT THAT STILL ISN'T ENOUGH. WE MUST ALSO LEARN TO DO THE RIGHT THING. THAT IS WHERE THE NEXT GREAT ADVANCE IN SOFTWARE-'KNOWLEDGEWARE'-COMES IN. UNTIL NOW, WE HAVE PROGRAMMED THE USER TO EXPLOIT THE MACHINE. IN THE FUTURE, WE MUST REVERSE THE PROCESS-PROGRAM THE MACHINE TO EXPLOIT THE POTENTIAL OF THE MIND. Desc.: TECHNOLOGICAL FORECASTING; COMPUTER SOFTWARE; ARTIFICIAL INTELLIGENCE Ident.: ARTIFICIAL INTELLIGENCE; EXPERT SYSTEMS; GATEWAY SOFTWARE; COMMERCIAL DATABASES; KNOWLEDGEWARE Class Codes: C6000; C1230 Y034013 129 1309843 C84037547 DISTRIBUTED ARCHITECTURE AND PARALLEL NONDIRECTIONAL SEARCH FOR KNOWLEDGE-BASED CARTOGRAPHIC FEATURE EXTRACTION SYSTEMS LAMBIRD, B.A.; LAVINE, D.; KANAL, L.N. LNK CORP., SILVER SPRING, MD, USA INT. J. MAN-MACH. STUD. (GB) VOL.20, NO.1 107-20 JAN. 1984 CODEN: IJMMBC ISSN: 0020-7373 U. S. Copyright Clearance Center Code: 0020-7373/84/010107+14$03.00/0 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (20 Refs) EXPERT OR KNOWLEDGE-BASED SYSTEM APPROACHES ARE CURRENTLY BEING VIEWED WITH GREAT INTEREST FOR THEIR POTENTIAL TO HANDLE THE MANY DIFFICULT PROBLEMS ENCOUNTERED IN IMAGE UNDERSTANDING AND CARTOGRAPHIC FEATURE EXTRACTION FROM REMOTELY SENSED IMAGERY. THIS ARTICLE PRESENTS AN OVERVIEW OF THE MANY TYPES OF KNOWLEDGE THAT MUST BE MODELED IN REMOTE SENSING AND CARTOGRAPHY, AND DISCUSSES ARCHITECTURAL AND CONTROL ASPECTS DEEMED IMPORTANT FOR CARTOGRAPHIC EXPERT SYSTEMS. A DISTRIBUTED ARCHITECTURE AND A CONTROL STRUCTURE BASED ON A PARALLEL NONDIRECTIONAL SEARCH ALGORITHM ARE DESCRIBED AND OPEN PROBLEMS ARE MENTIONED. Desc.: PICTURE PROCESSING; REMOTE SENSING; ARTIFICIAL INTELLIGENCE CARTOGRAPHY Ident.: PARALLEL NONDIRECTIONAL SEARCH; IMAGE UNDERSTANDING; CARTOGRAPHIC FEATURE EXTRACTION; REMOTELY SENSED IMAGERY; REMOTE SENSING; CARTOGRAPHY; CARTOGRAPHIC EXPERT SYSTEMS; DISTRIBUTED ARCHITECTURE Class Codes: C1230; C1250 Y034013 130 1297765 C84036806 USING CHUNKING TO SOLVE CHESS PAWN ENDGAMES BERLINER, H.; CAMPBELL, M. DEPT. OF COMPUTER SCI., CARNEGIE-MELLON UNIV., PITTSBURGH, PA, USA ARTIF. INTELL. (NETHERLANDS) VOL.23, NO.1 97-120 MAY 1984 CODEN: AINTBB ISSN: 0004-3702 U. S. Copyright Clearance Center Code: 0004-3702/84/$3.00 Treatment: PRACTICAL Doc. Type: JRNL PAP. Lang.: ENG. (14 Refs) CHUNKER IS A CHESS PROGRAM THAT USES CHUNKED KNOWLEDGE TO IMPROVE ITS PERFORMANCE. ITS DOMAIN IS A SUBSET OF KING AND PAWN ENDINGS IN CHESS THAT HAS BEEN STUDIED FOR OVER 300 YEARS. CHUNKER HAS A LARGE LIBRARY OF CHUNK INSTANCES WHERE EACH CHUNK TYPE HAS A PROPERTY LIST AND EACH INSTANCE HAS A SET OF VALUES FOR THESE PROPERTIES. THIS ALLOWS CHUNKER TO REASON ABOUT POSITIONS THAT COME UP IN THE SEARCH THAT WOULD OTHERWISE HAVE TO BE HANDLED BY MEANS OF ADDITIONAL SEARCH. THUS THE PROGRAM IS ABLE TO SOLVE THE MOST DIFFICULT PROBLEM OF ITS PRESENT DOMAIN (A PROBLEM THAT WOULD REQUIRE 45 PLY OF SEARCH AND ON THE ORDER OF 10/SUP 13/ YEARS OF CPU TIME TO BE SOLVED BY THE BEST OF PRESENT-DAY CHESS PROGRAMS) IN 18 PLY AND ONE MINUTE OF CPU TIME. FURTHER, CHUNKER IS UNDOUBTEDLY THE WORLD'S FOREMOST EXPERT IN ITS DOMAIN, AND HAS DISCOVERED TWO MISTAKES IN THE LITERATURE AND HAS BEEN INSTRUMENTAL IN DISCOVERING A NEW THEOREM ABOUT THE DOMAIN THAT ALLOWS THE ASSESSING OF POSITIONS WITH NEW DEGREE OF EASE AND CONFIDENCE. IN THIS PAPER THE AUTHORS SHOW HOW THE LIBRARIES ARE COMPILED, HOW CHUNKER WORKS, AND DISCUSSES PLANS FOR EXTENDING IT TO PLAY THE WHOLE DOMAIN OF KING AND PAWN ENDINGS. Desc.: GAMES OF SKILL Ident.: GAMES OF SKILL; CHESS PAWN ENDGAMES; CHUNKER; CHESS PROGRAM CHUNKED KNOWLEDGE Class Codes: C7890; C1230 Y034013 131 1294818 C84032608 ON GAME GRAPH STRUCTURE AND ITS INFLUENCE ON PATHOLOGY NAU, D.S. DEPT. OF COMPUTER SCI., UNIV. OF MARYLAND, COLLEGE PARK, MD, USA INT. J. COMPUT. AND INF. SCI. (USA) VOL.12, NO.6 367-83 DEC. 1983 CODEN: IJCIAH ISSN: 0091-7036 U. S. Copyright Clearance Center Code: 0091-7036/83/1200-0367$03.00/0 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (22 Refs) ALMOST ALL GAME TREE SEARCH PROCEDURES USED IN ARTIFICIAL INTELLIGENCE ARE VARIANTS ON MINIMAXING. UNTIL RECENTLY, IT WAS ALMOST UNIVERSALLY BELIEVED THAT SEARCHING DEEPER ON GAME TREES WITH SUCH PROCEDURES WOULD IN GENERAL YIELD A BETTER DECISION. HOWEVER, RECENT INVESTIGATIONS SHOW THAT THERE ARE MANY 'PATHOLOGICAL' GAME TREES FOR WHICH SEARCHING DEEPER CONSISTENTLY DEGRADES THE DECISION. THIS PAPER INVESTIGATES ONE POSSIBLE CAUSE OF PATHOLOGY. IN PARTICULAR, A CLASS OF GAMES THAT IS NORMALLY PATHOLOGICAL IS SHOWN TO BECOME NONPATHOLOGICAL WHEN THE GAMES ARE MODIFIED SO THAT GAME POSITIONS CAN BE REACHED BY MORE THAN ONE PATH. THIS RESULT SUGGESTS THAT IN GENERAL, PATHOLOGY IS LESS LIKELY WHEN GAME POSITIONS CAN BE REACHED BY MORE THAN ONE PATH. THIS MAY BE ONE REASON WHY GAMES SUCH AS CHESS AND CHECKERS ARE NONPATHOLOGICAL. IN ADDITION, THIS RESULT SUPPORTS THE HYPOTHESIS THAT PATHOLOGY IS LESS LIKELY WHEN SIBLING NODES HAVE SIMILAR MINIMAX VALUES. THIS PAPER ALSO INVESTIGATES A POSSIBLE CURE FOR PATHOLOGY-AN ALTERNATIVE TO MINIMAXING CALLED PROBABILITY ESTIMATION WHICH HAS BEEN SHOWN TO AVOID PATHOLOGY AND THUS PRODUCE MORE ACCURATE DECISIONS THAN MINIMAXING ON AT LEAST ONE PATHOLOGICAL GAME. Desc.: GAME THEORY; ARTIFICIAL INTELLIGENCE; GRAPH THEORY Ident.: GAME GRAPH STRUCTURE; PATHOLOGY; GAME TREE SEARCH PROCEDURES ; ARTIFICIAL INTELLIGENCE; MINIMAXING; SIBLING NODES Class Codes: C1230; C1140E; C1160 Y034013 132 1294815 C84032605 ON THE APPLICATION OF RULE-BASED TECHNIQUES TO THE DESIGN OF ADVICE-GIVING SYSTEMS JACKSON, P.; LEFRERE, P. INST. OF EDUCATIONAL TECHNOL., OPEN UNIV., MILTON KEYNES, ENGLAND INT. J. MAN-MACH. STUD. (GB) VOL.20, NO.1 63-86 JAN. 1984 CODEN: IJMMBC ISSN: 0020-7373 U. S. Copyright Clearance Center Code: 0020-7373/84/010063+24$03.00/0 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (69 Refs) THIS ARTICLE ATTEMPTS TO ASSESS HOW MUCH IS KNOWN ABOUT BUILDING SYSTEMS WHOSE ADVICE ACTUALLY BENEFITS USERS. THE AUTHORS REVIEW CURRENT APPROACHES TO THE PROVISION OF ONLINE HELP, AND SUGGEST THAT THE MOST PROMISING ARE THOSE WHICH REPRESENT A USER'S INTENTIONS EXPLICITLY. FOLLOWING THIS LEAD, THEY EXAMINE RECENT WORK ON SPEECH ACTS, PLANNING AND META-LEVEL INFERENCE FOR CLUES AS TO HOW A USER'S INPUTS COULD BE INTERPRETED IN THE CONTEXT OF HIS CURRENT AIMS AND ACTIVITIES. THEY CONCLUDE THAT THE APPROPRIATE CONTEXT OF INTERPRETATION FOR AN INPUT IS SUPPLIED BY HYPOTHESES CONCERNING THE CURRENT STATE OF A USER'S PLAN. NEXT THEY SUGGEST THAT THE TECHNIQUES DEVELOPED IN RULE-BASED SYSTEMS COULD BE USED TO IMPLEMENT AN ADVISER CAPABLE OF GENERATING AND REVISING PLAN HYPOTHESES, AND OUTLINE WHAT THEY CONSIDER TO BE THE OUTSTANDING PROBLEMS OF CONTROL ASSOCIATED WITH SUCH AN IMPLEMENTATION. FINALLY, THEY SHOW HOW SUCH A SYSTEM MIGHT HELP A USER TO STRUCTURE HIS ACTIVITY, SO THAT HE CAN ITERATE TOWARDS HIS GOAL WHILE AVOIDING COMMON ERRORS. Desc.: ARTIFICIAL INTELLIGENCE; MAN-MACHINE SYSTEMS Ident.: EXPERT SYSTEMS; ADVICE-GIVING SYSTEMS; ONLINE HELP; SPEECH ACTS; PLANNING; META-LEVEL INFERENCE; RULE-BASED SYSTEMS Class Codes: C1230; C1270 Y034013 133 1294814 C84032604 KNOWLEDGE REORGANIZATION AND REASONING STYLE RIESBECK, C.K. DEPT. OF COMPUTER SCI., YALE UNIV., NEW HAVEN, CT, USA INT. J. MAN-MACH. STUD. (GB) VOL.20, NO.1 45-61 JAN. 1984 CODEN: IJMMBC ISSN: 0020-7373 U. S. Copyright Clearance Center Code: 0020-7373/84/010045+17$03.00/0 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (23 Refs) TO STUDY THE LEARNING OF EXPERTISE, TWO CLOSELY RELATED STAGES OF EXPERTISE IN ECONOMICS REASONING ARE ANALYZED AND MODELLED, AND A MECHANISM FOR GOING FROM THE FIRST TO THE SECOND IS PROPOSED. BOTH STAGES SHARE THE SAME BASIC CONCEPTS AND GENERATE PLAUSIBLE ECONOMIC SCENARIOS, BUT REASONING IN THE FIRST STAGE OVERSIMPLIFIES BY FOCUSING ON HOW THE GOALS OF A FEW ACTORS ARE AFFECTED. REASONING IN THE SECOND STAGE PRODUCES BETTER ARGUMENTS BY TAKING INTO ACCOUNT HOW ALL THE RELEVANT PARTS OF THE ECONOMY MIGHT BE AFFECTED. THE FIRST STAGE IS MODELLED BY HIGHLY INTERCONNECTED GOAL FORESTS AND VERY SELECTIVE STORY UNDERSTANDING SEARCH HEURISTICS. THE SECOND STAGE IS MODELLED WITH MORE EXPLICIT LINKS BETWEEN ECONOMIC QUANTITIES AND A MORE APPROPRIATE SET OF SEARCH HEURISTICS. THE LEARNING MECHANISM IS A FAILURE-DRIVEN PROCESS THAT NOT ONLY RECORDS BETTER ARGUMENTS AS THEY ARE SEEN, BUT ALSO RECORDS THE FAILURE OF EXISTING INFERENCE RULES TO FIND THESE ARGUMENTS ON THEIR OWN. THE COLLECTED FAILURES ARE USED TO DETERMINE WHICH SEARCH HEURISTICS WORK BEST IN WHICH SITUATIONS. Desc.: ARTIFICIAL INTELLIGENCE Ident.: EXPERT SYSTEMS; KNOWLEDGE REORGANISATION; ECONOMICS REASONING; REASONING; GOAL FORESTS; STORY UNDERSTANDING SEARCH HEURISTICS; ECONOMIC QUANTITIES; LEARNING MECHANISM Class Codes: C1230 Y034013 134 1294813 C84032603 EXPERT SYSTEMS: AN ALTERNATIVE PARADIGM COOMBS, M.; ALTY, J. DEPT. OF COMPUTER SCI., UNIV. OF STRATHCLYDE, GLASGOW, SCOTLAND INT. J. MAN-MACH. STUD. (GB) VOL.20, NO.1 21-43 JAN. 1984 CODEN: IJMMBC ISSN: 0020-7373 U. S. Copyright Clearance Center Code: 0020-7373/84/010021+23$03.00/0 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (30 Refs) THERE HAS RECENTLY BEEN A SIGNIFICANT EFFORT BY THE AI COMMUNITY TO INTEREST INDUSTRY IN THE POTENTIAL OF EXPERT SYSTEMS. HOWEVER, THIS HAS RESULTED IN FAR FEWER SUBSTANTIAL APPLICATIONS PROJECTS THAN MIGHT BE EXPECTED. THIS ARTICLE ARGUES THAT THIS IS BECAUSE HUMAN EXPERTS ARE RARELY REQUIRED TO PERFORM THE ROLE THAT COMPUTER-BASED EXPERTS ARE PROGRAMMED TO ADOPT. INSTEAD OF BEING CALLED IN TO ANSWER WELL-DEFINED PROBLEMS, THEY ARE MORE OFTEN ASKED TO ASSIST OTHER EXPERTS TO EXTEND AND REFINE THEIR UNDERSTANDING OF A PROBLEM AREA AT THE JUNCTION OF THEIR TWO DOMAINS OF KNOWLEDGE. THIS MORE PROPERLY INVOLVES EDUCATIONAL RATHER THAN PROBLEM-SOLVING SKILLS. AN ALTERNATIVE APPROACH TO EXPERT SYSTEM DESIGN IS PROPOSED BASED UPON GUIDED DISCOVERY LEARNING. THE USER IS PROVIDED WITH A SUPPORTIVE ENVIRONMENT FOR A PARTICULAR CLASS OF PROBLEM, THE SYSTEM PREDOMINANTLY ACTING AS AN ADVISER RATHER THAN DIRECTING THE INTERACTION. THE ENVIRONMENT INCLUDES A DATABASE OF DOMAIN KNOWLEDGE, A SET OF PROCEDURES FOR ITS APPLICATION TO A CONCRETE PROBLEM, AND AN INTELLIGENT MACHINE-BASED ADVISER TO JUDGE THE USER'S EFFECTIVENESS AND ADVISE ON STRATEGY. THE PROCEDURES FOCUS UPON THE USE OF USER GENERATED 'EXPLANATIONS' BOTH TO PROMOTE THE APPLICATION OF DOMAIN KNOWLEDGE AND TO EXPOSE UNDERSTANDING DIFFICULTIES. SIMPLE DATABASE PROLOG IS BEING USED AS THE SUBJECT MATERIAL FOR THE PROTOTYPE SYSTEM WHICH IS KNOWN AS MINDPAD. Desc.: ARTIFICIAL INTELLIGENCE Ident.: AI; EXPERT SYSTEMS; EXPERT SYSTEM DESIGN; GUIDED DISCOVERY LEARNING; DOMAIN KNOWLEDGE; MINDPAD Class Codes: C1230 Y034013 135 1294806 C84032596 GENERAL BRANCH AND BOUND, AND ITS RELATION TO A* AND AO* NAU, D.S.; KUMAR, V.; KANAL, L. COMPUTER SCI. DEPT., UNIV. OF MARYLAND COLLEGE PARK, MD, USA ARTIF. INTELL. (NETHERLANDS) VOL.23, NO.1 29-58 MAY 1984 CODEN: AINTBB ISSN: 0004-3702 U. S. Copyright Clearance Center Code: 0004-3702/84/$3.00 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (25 Refs) BRANCH AND BOUND (BANDB) IS A PROBLEM-SOLVING TECHNIQUE WHICH IS WIDELY USED FOR VARIOUS PROBLEMS ENCOUNTERED IN OPERATIONS RESEARCH AND COMBINATORIAL MATHEMATICS. VARIOUS HEURISTIC SEARCH PROCEDURES USED IN ARTIFICIAL INTELLIGENCE (AI) ARE CONSIDERED TO BE RELATED TO BANDB PROCEDURES. HOWEVER, IN THE ABSENCE OF ANY GENERALLY ACCEPTED TERMINOLOGY FOR BANDB PROCEDURES, THERE HAVE BEEN WIDELY DIFFERING OPINIONS REGARDING THE RELATIONSHIPS BETWEEN THESE PROCEDURES AND BANDB. THE PAPER PRESENTS A FORMULATION OF BANDB GENERAL ENOUGH TO INCLUDE PREVIOUS FORMULATIONS AS SPECIAL CASES, AND SHOWS HOW TWO WELL-KNOWN AI SEARCH PROCEDURES (A* AND AO*) ARE SPECIAL CASES OF THIS GENERAL FORMULATION. Desc.: ARTIFICIAL INTELLIGENCE; HEURISTIC PROGRAMMING; OPTIMISATION Ident.: BRANCH AND BOUND; A*; AO*; PROBLEM-SOLVING TECHNIQUE; HEURISTIC SEARCH PROCEDURES; ARTIFICIAL INTELLIGENCE Class Codes: C1230; C1180 Y034013 136 1294805 C84032595 A HEURISTIC SEARCH ALGORITHM WITH MODIFIABLE ESTIMATE MERO, L. COMPUTER AND AUTOMATION INST., HUNGARIAN ACAD. OF SCI., BUDAPEST, HUNGARY ARTIF. INTELL. (NETHERLANDS) VOL.23, NO.1 13-27 MAY 1984 CODEN: AINTBB ISSN: 0004-3702 U. S. Copyright Clearance Center Code: 0004-3702/84/$3.00 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (14 Refs) DESCRIBES AN IMPROVED VERSION OF TWO PREVIOUSLY PUBLISHED ALGORITHMS IN THE AREA: A* AND B. THE NEW APPROACH IS BASED ON CONSIDERING THE ESTIMATE H> (N) ON NODE N AS A VARIABLE RATHER THAN AS A CONSTANT. THE NEW ALGORITHM THUS IMPROVES THE ESTIMATE AS IT GOES ON, AVOIDING SOME USELESS NODE EXPANSIONS. IT IS PROVED TO NEVER EXPAND MORE NODES THAN B OR A* AND TO EXPAND A MUCH SMALLER NUMBER OF THEM IN SOME CASES. ANOTHER RESULT OF THE PAPER IS A PROOF THAT NO OVERALL OPTIMAL ALGORITHM EXISTS IF THE COST OF AN ALGORITHM IS MEASURED BY THE TOTAL NUMBER OF NODE EXPANSIONS. Desc.: HEURISTIC PROGRAMMING; OPTIMISATION Ident.: OPTIMISATION; A* ALGORITHM; B ALGORITHM; HEURISTIC PROGRAMMING; HEURISTIC SEARCH ALGORITHM; MODIFIABLE ESTIMATE; NODE EXPANSIONS Class Codes: C1230; C1180 Y034013 137 1294802 C84032592 DOMAIN-INDEPENDENT PLANNING: REPRESENTATION AND PLAN GENERATION WILKINS, D.E. ARTIFICIAL INTELLIGENCE CENTER, SRI INTERNATIONAL, MENLO PARK, CA, USA ARTIF. INTELL. (NETHERLANDS) VOL.22, NO.3 269-301 APRIL 1984 CODEN: AINTBB ISSN: 0004-3702 U. S. Copyright Clearance Center Code: 0004-3702/84/$3.00 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (17 Refs) A DOMAIN-INDEPENDENT PLANNING PROGRAM THAT SUPPORTS BOTH AUTOMATIC AND INTERACTIVE GENERATION OF HIERARCHICAL, PARTIALLY ORDERED PLANS IS DESCRIBED. AN IMPROVED FORMALISM MAKES EXTENSIVE USE OF CONSTRAINTS AND RESOURCES TO REPRESENT DOMAINS AND ACTIONS MORE POWERFULLY. THE FORMALISM ALSO OFFERS EFFICIENT METHODS FOR REPRESENTING PROPERTIES OF OBJECTS THAT DO NOT CHANGE OVER TIME, ALLOWS SPECIFICATION OF THE PLAN RATIONALE (WHICH INCLUDES SCOPING OF CONDITIONS AND APPROPRIATELY RELATING DIFFERENT LEVELS IN THE HIERARCHY), AND PROVIDES THE ABILITY TO EXPRESS DEDUCTIVE RULES FOR DEDUCING THE EFFECTS OF ACTIONS. THE IMPLICATIONS OF ALLOWING PARALLEL ACTIONS IN A PLAN OR PROBLEM SOLUTION ARE DISCUSSED, AND NEW TECHNIQUES FOR EFFICIENTLY DETECTING AND REMEDYING HARMFUL PARALLEL INTERACTIONS ARE PRESENTED. THE MOST IMPORTANT OF THESE TECHNIQUES, REASONING ABOUT RESOURCES, IS EMPHASIZED AND EXPLAINED. THE SYSTEM SUPPORTS CONCURRENT EXPLORATION OF DIFFERENT BRANCHES IN THE SEARCH, MAKING BEST-FIRST SEARCH EASY TO IMPLEMENT. Desc.: ARTIFICIAL INTELLIGENCE; ADMINISTRATIVE DATA PROCESSING Ident.: ARTIFICIAL INTELLIGENCE; AP; DOMAIN-INDEPENDENT PLANNING PROGRAM Class Codes: C1230; C7100 Y034013 138 1294656 C84032364 ABUNDANT APPLICATIONS BEGUILE NEW USERS BARNES, R. COMPUTERWORLD (USA) VOL.18, NO.18 SR/70-2 30 APRIL 1984 CODEN: CMPWAB ISSN: 0010-4841 Treatment: GENERAL, REVIEW Doc. Type: JRNL PAP. Lang.: ENG. THE PROLIFERATION OF MICROCOMPUTERS IN THE BUSINESS ENVIRONMENT IS BEING TRAILED CLOSELY BY A PROLIFERATION OF INEXPENSIVE PACKAGED APPLICATIONS. THE LOW PURCHASE PRICE OF THESE APPLICATIONS LULLS THE INEXPERIENCED COMPUTER USER INTO A FALSE SENSE OF ECONOMY AND SECURITY. ONLY WHEN THE REALITIES OF LEARNING COMPUTER PROGRAMMING AND OPERATING SYSTEM INTERNALS ARE DISCOVERED DOES THE BUYER BEGIN A SEARCH FOR A PACKAGE THAT SOMEONE ELSE HAS WRITTEN. Desc.: E.D.P. MANAGEMENT; COMPUTER APPLICATIONS; COMPUTER SOFTWARE Ident.: APPLICATION SOFTWARE PACKAGES; NEW USERS; MICROCOMPUTERS; BUSINESS ENVIRONMENT; PURCHASE PRICE; COMPUTER PROGRAMMING; OPERATING SYSTEM INTERNALS Class Codes: C0310F; C7000 Y034013 139 1280044 C84030908, D84001420 SOFTWARE. NEW GENRE PROMOTES BUSINESS AND MANAGEMENT EXPERTISE RUMMER, P. SMALL SYST. WORLD (USA) VOL.12, NO.5 65-6 MAY 1984 CODEN: SSWOD5 ISSN: 0272-5444 Treatment: GENERAL, REVIEW Doc. Type: JRNL PAP. Lang.: ENG. THIS SPRING HAS SEEN THE DEBUT OF SOFTWARE PACKAGES THAT CAN HELP AN END USER IMPROVE SPECIFIC BUSINESS AND PERSONAL MANAGEMENT SKILLS-FINANCIAL MANAGEMENT, SALES, MOTIVATION, AND WRITING, FOR EXAMPLE. THESE PROGRAMS ARE 'EXPERT' SOFTWARE IN THE SENSE THAT THEY RELY ON THE KNOWLEDGE OF PROFESSIONALS IN A SPECIFIC SUBJECT TO ASSESS THE USER'S NEEDS. IN GENERAL, MANAGEMENT DEVELOPMENT SOFTWARE FOR MICROS FALLS INTO TWO CATEGORIES: PROGRAMS THAT DEVELOP BUSINESS AND FINANCIAL MANAGEMENT SKILLS (E.G. OZ: MANAGEMENT CONTROL, DSS OPTIONWARE, AND SALESEYE), AND PROGRAMS THAT IMPROVE THE USER'S PERSONAL BUSINESS AND MANAGEMENT SKILLS (E.G. THE HUMAN EDGE, THOUGHTWARE, AND WRITING IS THINKING). Desc.: ADMINISTRATIVE DATA PROCESSING; ARTIFICIAL INTELLIGENCE Ident.: PERSONAL BUSINESS SKILLS; EXPERT SYSTEMS; EXPERTISE; SOFTWARE PACKAGES; PERSONAL MANAGEMENT SKILLS; SALES; MOTIVATION; WRITING; MANAGEMENT DEVELOPMENT SOFTWARE; FINANCIAL MANAGEMENT SKILLS; OZ: MANAGEMENT CONTROL; DSS OPTIONWARE; SALESEYE; THE HUMAN EDGE; THOUGHTWARE; WRITING IS THINKING Class Codes: C7100; D2010 Y034013 140 1279810 C84030671 DIALOS/INT-A SYSTEM WITH ELEMENTS OF ARTIFICIAL INTELLIGENCE FOR THE SOLUTION OF PROBLEM CASES KALJN, A.; SZPUNAR, Z. TECH. HOCHSCHULE 'OTTO VON GUERICKE', MAGDEBURG, GERMANY WISS. Z. TECH. HOCHSCH. OTTO VON GUERICKE MAGDEBURG (GERMANY) VOL.27, NO.8 41-3 1983 CODEN: WCGMAI ISSN: 0541-8933 Treatment: PRACTICAL Doc. Type: JRNL PAP. Lang.: GERMAN (3 Refs) THE SYSTEM DIALOS/INT WAS DEVELOPED AT THE WROCLAW POLYTECHNIC. THE COMMUNICATION LANGUAGE COMPRISES 41 KEYWORDS. AN APPLICATION TO THE SOLUTION OF A PROBLEM MUST CONSIST OF THREE STATEMENT LISTS; THE OUTPUT SITUATION S (P), THE TARGET SITUATION S (D) AND OPERATORS TO SWITCH EACH ITEM S (I) INTO AN ALTERNATIVE PATH S (J). FRAGMENTS ON A TYPICAL PROGRAM ARE GIVEN AND THE ORGANISATION OF THE SEARCH PROCESS IS DESCRIBED. THE PROGRAM MAY OPERATE IN EITHER CONVERSATIONAL OR JOB MODE, AND HAS BEEN RUN ON THE ODRA 1300 MACHINE USING THE OPERATING SYSTEM GEORGE 3. Desc.: HIGH LEVEL LANGUAGES; ARTIFICIAL INTELLIGENCE Ident.: ARTIFICIAL INTELLIGENCE; PROBLEM CASES; DIALOS/INT; COMMUNICATION LANGUAGE; OUTPUT SITUATION; TARGET SITUATION; OPERATORS; SEARCH PROCESS Class Codes: C6140D; C1230 Y034013 141 1279299 C84030006 OUTLINE OF THE PERSONAL SEQUENTIAL INFERENCE MACHINE: PSI UCHIDA, S.; YOKOTA, M.; YAMAMOTO, A.; TAKI, K.; NISHIKAWA, H. ICOT RES. CENTER, INST. FOR NEW GENERATION COMPUTER TECHNOL. TOKYO, JAPAN NEW GENERATION COMPUT. (JAPAN) VOL.1, NO.1 75-9 1983 CODEN: NGCOE5 ISSN: 0288-3635 Treatment: PRACTICAL Doc. Type: JRNL PAP. Lang.: ENG. (4 Refs) PSI IS A PERSONAL COMPUTER SYSTEM BEING DEVELOPED AS A TOOL FOR PROVIDING RESEARCHERS WITH AN EFFICIENT PROGRAMMING ENVIRONMENT. IT DIRECTLY SUPPORTS LOGIC PROGRAMMING LANGUAGES, KL0 (FIFTH GENERATION KERNEL LANGUAGE, VERSION 0), WITH FIRMWARE AND HARDWARE. ITS INTERPRETER IS IMPLEMENTED IN THE FIRMWARE AND SEVERAL HARDWARE MECHANISMS ARE PROVIDED TO ATTAIN ALMOST THE SAME LEVEL OF PERFORMANCE AS THE DEC-10 PROLOG ON DEC2060. IT ALSO PROVIDES THE USER WITH A LARGE MEMORY SPACE, 40 BITS*2 TO 16 MW, WHICH IS ESSENTIAL FOR DEVELOPING ACTUAL APPLICATION PROGRAMS LIKE AN EXPERT SYSTEM. TO MAKE EFFICIENT MAN-MACHINE INTERACTION POSSIBLE, SUCH INPUT AND OUTPUT DEVICES AS BIT-MAP DISPLAY, POINTING DEVICE AND KEYBOARD ARE PROVIDED. A LOCAL AREA NETWORK IS ALSO BEING DEVELOPED TO BUILD A DISTRIBUTED SYSTEM. Desc.: MICROCOMPUTERS Ident.: PERSONAL SEQUENTIAL INFERENCE MACHINE; PSI; PERSONAL COMPUTER SYSTEM; PROGRAMMING ENVIRONMENT; LOGIC PROGRAMMING LANGUAGES; KL0 INTERPRETER; DEC-10 PROLOG; DEC2060; APPLICATION PROGRAMS; EXPERT SYSTEM; MAN-MACHINE INTERACTION; BIT-MAP DISPLAY; POINTING DEVICE; LOCAL AREA NETWORK Class Codes: C5410 Y034013 142 1279102 C84029607 CASE ROLE FILLING AS A SIDE EFFECT OF VISUAL SEARCH MARBURGER, H.; WAHLSTER, W. RES. UNIT FOR INFORMATION SCI. AND ARTIFICIAL INTELLIGENCE, UNIV. OF HAMBURG, HAMBURG, GERMANY FIRST CONFERENCE OF THE EUROPEAN CHAPTER OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS. PROCEEDINGS OF THE CONFERENCE 188-95 1983 1-2 SEPT. 1983 PISA, ITALY Publ: ASSOC. COMPUT. LINGUISTICS, MENLO PARK, CA, USA VIII+203 pp. Treatment: THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (13 Refs) THIS PAPER ADDRESSES THE PROBLEM OF GENERATING COMMUNICATIVELY ADEQUATE EXTENDED RESPONSES IN THE ABSENCE OF SPECIFIC KNOWLEDGE CONCERNING THE INTENTIONS OF THE QUESTIONER. THE AUTHORS FORMULATE AND JUSTIFY A HEURISTIC FOR THE SELECTION OF OPTIONAL DEEP CASE SLOTS NOT CONTAINED IN THE QUESTION AS CANDIDATES FOR THE ADDITIONAL INFORMATION CONTAINED IN AN EXTENDED RESPONSE. IT IS SHOWN THAT, IN A VISUALLY PRESENT DOMAIN OF DISCOURSE, CASE ROLE FILLING FOR THE CONSTRUCTION OF AN EXTENDED RESPONSE CAN BE REGARDED AS A SIDE EFFECT OF THE VISUAL SEARCH NECESSARY TO ANSWER A QUESTION CONTAINING A LOCOMOTION VERB. THE PAPER DESCRIBES THE VARIOUS REPRESENTATION CONSTRUCTIONS USED IN THE GERMAN LANGUAGE DIALOG SYSTEM HAM-ANS FOR DEALING WITH THE SEMANTICS OF LOCOMOTION VERBS AND ILLUSTRATES THEIR USE IN GENERATING EXTENDED RESPONSES. IN PARTICULAR, IT OUTLINES THE STRUCTURE OF THE GEOMETRICAL SCENE DESCRIPTION, THE REPRESENTATION OF EVENTS IN A LOGIC-ORIENTED SEMANTIC REPRESENTATION LANGUAGE, THE CASE-FRAME LEXICON AND THE REPRESENTATION OF THE REFERENTIAL SEMANTICS BASED ON THE FLAVOR SYSTEM. THE EMPHASIS IS ON A DETAILED PRESENTATION OF THE APPLICATION OF OBJECT-ORIENTED PROGRAMMING METHODS FOR COPING WITH THE SEMANTICS OF LOCOMOTION VERBS. THE PROCESS OF GENERATING AN EXTENDED RESPONSE IS ILLUSTRATED BY AN EXTENSIVELY ANNOTATED TRACE. Desc.: COMPUTATIONAL LINGUISTICS; ARTIFICIAL INTELLIGENCE Ident.: DEEP CASE SLOTS; EXTENDED RESPONSE; CASE ROLE FILLING; VISUAL SEARCH; LOCOMOTION VERB; REPRESENTATION CONSTRUCTIONS; GERMAN LANGUAGE DIALOG SYSTEM; HAM-ANS; SEMANTICS; LOCOMOTION VERBS; GEOMETRICAL SCENE DESCRIPTION; SEMANTIC REPRESENTATION LANGUAGE; CASE-FRAME LEXICON; REFERENTIAL SEMANTICS; FLAVOR SYSTEM; OBJECT-ORIENTED PROGRAMMING Class Codes: C4290; C1230 Y034013 143 1278433 C84028494 HIERARCHICAL GOAL-SEEKING NEURAL ROBOTIC MODEL LIGOMENIDES, P.A. ELECTRICAL ENGNG. DEPT., UNIV. OF MARYLAND, COLLEGE PARK, MD, USA SINENCIO, E.S. (Editors) PROCEEDINGS OF THE 26TH MIDWEST SYMPOSIUM ON CIRCUITS AND SYSTEMS 62-5 1983 15-16 AUG. 1983 PUEBLA, MEXICO Publ: WESTERN PERIODICALS, NORTH HOLLYWOOD, CA, USA XIX+628 pp. Treatment: THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (12 Refs) THE EC MODEL, INTRODUCED BY LIGOMENIDES (SEE INT. J. POLICY ANAL. AND INF. SYST., VOL.6, NO.3, P.273-84, SEPT. 1982), IS PROPOSED FOR THE IMPLEMENTATION OF DECENTRALIZED GOAL-SEEKING INTELLIGENCE IN NEUTRAL ROBOTS. HOWEVER INCOMPLETE AS IT MAY BE IN ITS PRESENT FORM, ANALYSIS HAS SHOWN THAT THE EC MODEL MAY BE A SIGNIFICANT STEP FORWARD IN THE SEARCH FOR REALIZABLE ENGINEERING SOLUTION TO THE MICROCOMPUTER/VLSI BASED IMPLEMENTATION OF NEURAL ROBOTS. IT IS ALSO PROPOSED THAT THE EC MODEL MAY PROVIDE A SIMULATION TOOL FOR THE ANALYSIS OF STABILITY AND SELF-REPAIR PROPERTIES OF COMPLEX HUMAN CYBERNETIC AND NEURAL ROBOTIC SYSTEMS. Desc.: ARTIFICIAL INTELLIGENCE; ROBOTS; HIERARCHICAL SYSTEMS Ident.: HIERARCHICAL GOAL-SEEKING NEURAL ROBOTIC MODEL; ENGINEERING CYBERNETIC MODEL; EC MODEL; IMPLEMENTATION; DECENTRALIZED GOAL-SEEKING INTELLIGENCE; NEUTRAL ROBOTS; REALIZABLE ENGINEERING SOLUTION; MICROCOMPUTER/VLSI BASED IMPLEMENTATION; SIMULATION TOOL; STABILITY; SELF-REPAIR PROPERTIES; COMPLEX HUMAN CYBERNETIC; NEURAL ROBOTIC SYSTEMS Class Codes: C1230; C1340B Y034013 144 1278431 C84028490 NEW INSIGHTS INTO THE RELATIONSHIPS AMONG BRANCH AND BOUND, DYNAMIC PROGRAMMING AND HEURISTIC SEARCH PROCEDURES KANAL, L.N.; KUMAR, V. DEPT. OF COMPUTER SCI., UNIV. OF MARYLAND, COLLEGE PARK, MD, USA Sponsor: IEEE 1983 IEEE INTERNATIONAL SYMPOSIUM ON INFORMATION THEORY. ABSTRACTS OF PAPERS 36-7 1983 26-30 SEPT. 1983 ST. JOVITE, QUE., CANADA Publ: IEEE, NEW YORK, USA IV+166 pp. Treatment: THEORETICAL Doc. Type: CONFERENCE PAPER Lang.: ENG. (5 Refs) SUMMARY FORM ONLY GIVEN, AS FOLLOWS. THE AUTHORS FIRST CITE THE GREAT CONFUSION WHICH EXISTS IN THE LITERATURE REGARDING THE PRECISE NATURE OF DYNAMIC PROGRAMMING (DP), BRANCH AND BOUND (BANDB) PROCEDURES, THE RELATIONSHIPS BETWEEN THEM, AND THEIR RELATIONSHIPS TO VARIOUS HEURISTIC SEARCH PROCEDURES OF ARTIFICIAL INTELLIGENCE. THEY THEN BRIEFLY OUTLINE THE CONCEPT OF A COMPOSITE DECISION PROCESS AS A GENERAL MODEL FOR DISCRETE OPTIMIZATION PROBLEMS. THIS NEW MODEL IS APPLICABLE TO STATE-SPACE SEARCH, PROBLEM-REDUCTION SEARCH, GAME-TREE SEARCH VARIOUS DISCRETE DETERMINISTIC OPTIMIZATION PROBLEMS. IT HAS HELPED IN EXPLAINING THE RELATIONSHIPS AMONG VARIOUS SEARCH ALGORITHMS, IN PROVIDING INSIGHTS INTO THEIR INDIVIDUAL PECULIARITIES, AND IN SUGGESTING ADDITIONAL SEARCH PROCEDURES. SOME OF THE AUTHORS' RESULTS CONCERNING THE NATURE OF DP, BANDB, AND THEIR RELATIONSHIPS WITH EACH OTHER AND WITH VARIOUS HEURISTIC SEARCH PROCEDURES ARE SUMMARIZED, AND SOME OF THE CONFUSION CONCERNING DP, BANDB, AND HEURISTIC SEARCH ALGORITHMS SUCH AS A*, AO*, B*, SSS*, ALPHA-BETA AND THEIR GENERALIZATIONS ARE DISPELLED. Desc.: DYNAMIC PROGRAMMING; HEURISTIC PROGRAMMING; ARTIFICIAL INTELLIGENCE Ident.: BRANCH AND BOUND PROCEDURES; DYNAMIC PROGRAMMING; HEURISTIC SEARCH PROCEDURES; ARTIFICIAL INTELLIGENCE; COMPOSITE DECISION PROCESS; DISCRETE OPTIMIZATION; STATE-SPACE SEARCH; PROBLEM-REDUCTION SEARCH; GAME-TREE SEARCH; DETERMINISTIC OPTIMIZATION; A*; AO*; B*; SSS*; ALPHA-BETA Class Codes: C1230; C1180 Y034013 145 1278422 C84028481 ARTIFICIAL INTELLIGENCE-AN ADVANCE FOR THE MICROCOMPUTER ROHDE, U. MIKROCOMPUT. Z. (GERMANY) NO.4 74-7 APRIL 1984 CODEN: MDMZDL ISSN: 0720-4442 Treatment: PRACTICAL; THEORETICAL Doc. Type: JRNL PAP. Lang.: GERMAN THE AUTHOR STATES THAT BY SIMPLE SEARCHING A SMALL FRACTION OF INTELLIGENCE CAN BE SIMULATED. BEFORE IT REALLY COMES TO PROGRAMMING, WHICH SEEKS TO PROVE SOMETHING, FURTHERMORE TECHNICAL PREPARATIONS MUST BE MADE. A SPIN OFF IS THAT ONE CAN SEE WITH WHAT PRIMITIVE MEANS A SMALL BUT EXTENDABLE COMPILER CAN BE ESTABLISHED. HE POSES THE ARGUMENT THAT THE PROGRAM LANGUAGE TRIVIA IS LACKING IN SUBSTANCE AND COMMANDS MUST BE INTRODUCED INTO ROUTINES WHICH EFFICIENTLY MANIPULATE DATA. Desc.: ARTIFICIAL INTELLIGENCE Ident.: ARTIFICIAL INTELLIGENCE; MICROCOMPUTER; TECHNICAL PREPARATIONS; TRIVIA Class Codes: C1230 Y034013 146 1245009 C84020026 NEURONLIKE ADAPTIVE ELEMENTS THAT CAN SOLVE DIFFICULT LEARNING CONTROL PROBLEMS BARTO, A.G.; SUTTON, R.S.; ANDERSON, C.W. DEPT. OF COMPUTER AND INFORMATION SCI., UNIV. OF MASSACHUSETTS, AMHERST, MA, USA IEEE TRANS. SYST., MAN AND CYBERN. (USA) VOL.SMC-13, NO.5 834-46 SEPT.-OCT. 1983 CODEN: ISYMAW ISSN: 0018-9472 U. S. Copyright Clearance Center Code: 0018-9472/83/0900-0834$01.00 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (42 Refs) IT IS SHOWN HOW A SYSTEM CONSISTING OF TWO NEURONLIKE ADAPTIVE ELEMENTS CAN SOLVE A DIFFICULT LEARNING CONTROL PROBLEM. THE TASK IS TO BALANCE A POLE THAT IS HINGED TO A MOVABLE CART BY APPLYING FORCES TO THE CART'S BASE. IT IS ARGUED THAT THE LEARNING PROBLEMS FACED BY ADAPTIVE ELEMENTS THAT ARE COMPONENTS OF ADAPTIVE NETWORKS ARE AT LEAST AS DIFFICULT AS THIS VERSION OF THE POLE-BALANCING PROBLEM. THE LEARNING SYSTEM CONSISTS OF A SINGLE ASSOCIATIVE SEARCH ELEMENT (ASE) AND A SINGLE ADAPTIVE CRITIC ELEMENT (ACE). IN THE COURSE OF LEARNING TO BALANCE THE POLE, THE ASE CONSTRUCTS ASSOCIATIONS BETWEEN INPUT AND OUTPUT BY SEARCHING UNDER THE INFLUENCE OF REINFORCEMENT FEEDBACK, AND THE ACE CONSTRUCTS A MORE INFORMATIVE EVALUATION FUNCTION THAN REINFORCEMENT FEEDBACK ALONE CAN PROVIDE. THE DIFFERENCES BETWEEN THIS APPROACH AND OTHER ATTEMPTS TO SOLVE PROBLEMS USING NEUROLIKE ELEMENTS ARE DISCUSSED, AS IS THE RELATION OF THIS WORK TO CLASSICAL AND INSTRUMENTAL CONDITIONING IN ANIMAL LEARNING STUDIES AND ITS POSSIBLE IMPLICATIONS FOR RESEARCH IN THE NEUROSCIENCES. Desc.: ADAPTIVE CONTROL; LEARNING SYSTEMS; NEURAL NETS Ident.: NEURAL NETS; ADAPTIVE CONTROL; NEURONLIKE ADAPTIVE ELEMENTS LEARNING CONTROL PROBLEM; MOVABLE CART; ASSOCIATIVE SEARCH ELEMENT; ADAPTIVE CRITIC ELEMENT; ANIMAL LEARNING STUDIES Class Codes: C1240; C1230; C1340E; C1290L Y034013 147 1245005 C84020022 SOME RECENT RESULTS IN HEURISTIC SEARCH THEORY PEARL, J. DEPT. OF COMPUTER SCI., UNIV. OF CALIFORNIA, LOS ANGELES, CA, USA IEEE TRANS. PATTERN ANAL. AND MACH. INTELL. (USA) VOL.PAMI-6, NO.1 1-13 JAN. 1984 CODEN: ITPIDJ ISSN: 0162-8828 U. S. Copyright Clearance Center Code: 0162-8828/84/0100-0001$01.00 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (30 Refs) RECENT ANALYTICAL INVESTIGATIONS OF THE MATHEMATICAL PROPERTIES OF HEURISTICS AND THEIR INFLUENCE ON THE PERFORMANCE OF COMMON SEARCH TECHNIQUES ARE SUMMARIZED. THE RESULTS ARE REPORTED WITHOUT PROOFS, TOGETHER WITH DISCUSSIONS OF MOTIVATIONS AND INTERPRETATIONS. THE HIGHLIGHTS INCLUDE: THE OPTIMALITY OF A/SUP */; RELATIONS BETWEEN THE PRECISION OF THE HEURISTIC ESTIMATES AND THE AVERAGE COMPLEXITY OF THE SEARCH; COMPARISONS OF THE AVERAGE COMPLEXITIES OF A/SUP */ AND BACKTRACKING; PROCEDURES FOR COMPARING AND COMBINING NONADMISSIBLE HEURISTIC FUNCTIONS; THE INFLUENCE OF THE WEIGHT W (IN F= (L-W)G+WH) ON THE COMPLEXITY OF A/SUP */; THE PRUNING POWER OF ALPHABETA, SSS/SUP */, AND SCOUT; THE EFFECT OF SUCCESSOR ORDERING ON SEARCH COMPLEXITY, AND THE EFFECT OF SEARCH DEPTH OF THE QUALITY OF DECISIONS IN GAME PLAYING. Desc.: HEURISTIC PROGRAMMING Ident.: HEURISTIC SEARCH THEORY; OPTIMALITY; COMPLEXITY; BACKTRACKING; PRUNING; ALPHABETA; SSS/SUP */; SCOUT Class Codes: C1230 Y034013 148 1244997 C84020014 HOW MUCH IS CONTROL KNOWLEDGE WORTH? A PRIMITIVE EXAMPLE BARNETT, J.A. USC INFORMATION SCI. INST., MARINA DE REY, CA, USA ARTIF. INTELL. (NETHERLANDS) VOL.22, NO.1 77-89 JAN. 1984 CODEN: AINTBB ISSN: 0004-3702 U. S. Copyright Clearance Center Code: 0004-3702/84/$3.00 Treatment: THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (1 Refs) THE MOST BASIC ACTIVITY PERFORMED BY AN INTELLIGENT AGENT IS DECIDING WHAT TO DO NEXT. USUALLY, THE DECISION TAKES THE FORM OF SELECTING, FROM AMONG MANY APPLICABLE METHODS, ONE METHOD TO TRY FIRST, OR OPTING TO EXPAND A PARTICULAR NODE IN A SIMPLE SEARCH. THE MOST PRIMITIVE CASE IN SELECTING BETWEEN TWO INDEPENDENT ALTERNATIVES. THIS CASE IS EXAMINED AND THE VALUE OF THE CONTROL KNOWLEDGE THAT MAKES THE DECISION IS DETERMINED. ANOTHER RESULT DERIVED IS THE SENSITIVITY OF THE EXPECTED VALUE OF CONTROL KNOWLEDGE AS A FUNCTION OF THE ACCURACY OF THE PARAMETERS USED TO MAKE THESE CONTROL DECISIONS. Desc.: ARTIFICIAL INTELLIGENCE Ident.: CONTROL KNOWLEDGE; INTELLIGENT AGENT; CONTROL DECISIONS Class Codes: C1230 Y034013 149 1230457 C84018901 HUMAN-COMPUTER DIALOGUE DESIGN CONSIDERATIONS WILLIGES, R.C.; WILLIGES, B.H. DEPT. OF INDUSTRIAL ENGNG. AND OPERATIONS RES., VIRGINIA POLYTECH. INST. AND STATE UNIV., BLACKSBURG, VA, USA AUTOMATICA (GB) VOL.19, NO.6 767-73 NOV. 1983 CODEN: ATCAA9 ISSN: 0005-1098 U. S. Copyright Clearance Center Code: 0005-1098/83$3.00+0.0 Treatment: PRACTICAL; THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (24 Refs) THE THREE CRITICAL ELEMENTS OF QUALITY HUMAN-COMPUTER DIALOGUE DESIGN INCLUDE A COMPILATION OF EXISTING DESIGN GUIDELINES, BEHAVIORAL RESEARCH TO EXPAND THE GENERAL DIALOGUE PRINCIPLES, AND A COMPUTER-AIDED IMPLEMENTATION FOR RETRIEVAL OF THE GUIDELINES BY DIALOGUE AUTHORS. THE NEED FOR EMPIRICALLY BASED DIALOGUE GUIDELINES RELATING TO GENERIC HUMAN-COMPUTER TASKS IS STRESSED. THE RESULTS OF TWO STUDIES DEALING WITH INTERACTIVE TEXT EDITING ARE PRESENTED AS EXAMPLES OF RESEARCH DIRECTED TOWARD SPECIFYING DIALOGUE DESIGN PRINCIPLES AND DEVELOPING FORMAL TOOLS FOR HUMAN-COMPUTER DIALOGUE RESEARCH. ONE OF THESE STUDIES DEALS WITH A METHODOLOGY FOR DEVELOPING USER MODELS, AND THE SECOND STUDY DEALS WITH THE DESIGN OF HELP INFORMATION. COMPUTER-AIDING IN THE FORM OF RULE-BASED SYSTEMS FOR SELECTING APPROPRIATE DIALOGUE GUIDELINES AND SOFTWARE TOOLS FOR AUTHORING DIALOGUES ARE DISCUSSED AS MEANS FOR IMPLEMENTING THESE DIALOGUE DESIGN CONSIDERATIONS. Desc.: INTERACTIVE SYSTEMS; MAN-MACHINE SYSTEMS; HUMAN FACTORS; TEXT EDITING; CAD Ident.: CAD; MAN-MACHINE SYSTEMS; HUMAN FACTORS; HUMAN-COMPUTER DIALOGUE; DESIGN GUIDELINES; BEHAVIORAL RESEARCH; COMPUTER-AIDED IMPLEMENTATION; RETRIEVAL; HUMAN-COMPUTER TASKS; INTERACTIVE TEXT EDITING; HELP INFORMATION; SOFTWARE TOOLS Class Codes: C7400; C1270 Y034013 150 1230252 C84018656 AN EXPERT INTERMEDIARY SYSTEM FOR INTERACTIVE DOCUMENT RETRIEVAL GUIDA, G.; TASSO, C. ISTITUTO DI MATEMATICA, UNIV. DI UDINE, UDINE, ITALY AUTOMATICA (GB) VOL.19, NO.6 759-66 NOV. 1983 CODEN: ATCAA9 ISSN: 0005-1098 U. S. Copyright Clearance Center Code: 0005-1098/83$3.00+0.00 Treatment: PRACTICAL; THEORETICAL Doc. Type: JRNL PAP. Lang.: ENG. (18 Refs) CONSTRUCTING NATURAL LANGUAGE INTERFACES TO COMPUTER SYSTEMS OFTEN REQUIRES ACHIEVEMENT OF ADVANCED REASONING AND EXPERT CAPABILITIES IN ADDITION TO BASIC NATURAL LANGUAGE UNDERSTANDING. IN THE PAPER THE ABOVE ISSUE IS TACKLED IN THE FRAME OF AN ACTUAL APPLICATION CONCERNING THE DESIGN OF A NATURAL LANGUAGE INTERFACE FOR INTERACTIVE DOCUMENT RETRIEVAL. AFTER A SHORT DISCUSSION OF THE PECULIARITIES OF THIS APPLICATION, WHICH REQUIRES BOTH NATURAL LANGUAGE UNDERSTANDING AND REASONING CAPABILITIES, THE GENERAL ARCHITECTURE AND FUNDAMENTAL DESIGN CRITERIA OF A SYSTEM PRESENTLY BEING DEVELOPED AT THE UNIVERSITY OF UDINE ARE PRESENTED. THE SYSTEM, NAMED IR-NLI, IS AIMED AT ALLOWING NON-TECHNICAL USERS TO DIRECTLY ACCESS THROUGH NATURAL LANGUAGE THE SERVICES OFFERED BY ONLINE DATABASES. ATTENTION IS LATER FOCUSED ON THE BASIC FUNCTIONS OF IR-NLI, NAMELY UNDERSTANDING, DIALOGUE AND REASONING. AN EXAMPLE OF INTERACTION WITH IR-NLI IS FULLY WORKED OUT TO INTRODUCE THE MAIN FEATURES OF THE SYSTEM. KNOWLEDGE REPRESENTATION METHODS AND ALGORITHMS ADOPTED ARE THEN ILLUSTRATED. PERSPECTIVES AND DIRECTION FOR FUTURE RESEARCH ARE ALSO DISCUSSED. Desc.: INTERACTIVE SYSTEMS; ARTIFICIAL INTELLIGENCE; INFORMATION RETRIEVAL SYSTEMS Ident.: ARTIFICIAL INTELLIGENCE; EXPERT SYSTEMS; INFORMATION RETRIEVAL SYSTEMS; INTERACTIVE SYSTEMS; DOCUMENT RETRIEVAL; NATURAL LANGUAGE INTERFACES; COMPUTER SYSTEMS; ADVANCED REASONING; UNIVERSITY OF UDINE; IR-NLI; ONLINE DATABASES Class Codes: C7250; C1230