IRList Digest Tuesday, 7 May 1988 Volume 4 : Issue 33 Today's Topics: Abstract - Selected abstracts appearing in SIGIR FORUM (part 1 of 2) News addresses are Internet or CSNET: fox@vtopus.cs.vt.edu BITNET: foxea@vtvax3.bitnet ---------------------------------------------------------------------- Date: Tue, 17 May 88 09:10:51 CDT From: "Dr. Raghavan" Subject: Abstracts from SIGIR Forum [Part I of II - Ed.] Ed, These are the abstracts I included in the recent Forum. ... Regards, Vijay ABSTRACTS (Chosen by G. Salton from recent issues of journals in the retrieval area.) INFORMATION RETRIEVAL BY CONSTRAINED SPREADING ACTIVATION IN SEMANTIC NETWORKS Paul R. Cohen and Rick Kjeldsen, Department of Computer Information Science, Lederle Graduate Research Center, University of Massachusetts, Amherst, MA 01003 GRANT is an expert system for finding sources of funding given research proposals. Its search method - constrained spreading activation - makes inferences about the goals of the user and thus finds information that the user did not explicitly request but that is likely to be useful. The architecture of GRANT and the implementation of constrained spreading activation are described, and GRANT's performance is evaluated. (INFORMATION PROCESSING & MANAGEMENT, Vol. 23, No. 4, pp. 255-268, 1987) DEVELOPMENT OF THE CODER SYSTEM: A TESTBED FOR ARTIFICIAL INTELLIGENCE METHODS IN INFORMATION RETRIEVAL Edward A. Fox, Department of Computer Science, Virginia Tech, Blacksburg, VA 24061 The CODER (COmposite Document Expert/Extended/Effective Retrieval) system is testbed for investigating the application of artificial intelligence methods to increase the effectiveness of information retrieval systems. Particular attention is being given to analysis and representation of heterogeneous documents, such as electronic mail digests or messages, which vary widely in style, length, topic, and structure. Since handling passages of various types in these collections is difficult even for experimental systems like SMART, it is necessary to turn to other techniques being explored by information retrieval and artificial intelligence researchers. The CODER system architecture involves communities of experts around active blackboards, accessing knowledge bases that describe users, documents, and lexical items of various types. The initial lexical knowledge base construction work is now complete, and experts for search and time/date handling can perform a variety of processing tasks. User information and queries are being gathered, and a simple distributed skeletal system is operational. It appears that a number of artificial intelligence techniques are needed to best handle such common but complex document analysis and retrieval tasks. (INFORMATION PROCESSING & MANAGEMENT, Vol. 23, No. 4, pp. 341-366, 1987) USER MODELING IN INTELLIGENT INFORMATION RETRIEVAL Giorgio Brajnik, Giovanni Guida, and Carlo Tassom, Laboratorio di Intelligenza Artificiale, Dipartimento di Matematica e Informatica, Universita di Udine, Udine, Italy The issue of exploiting user modeling techniques in the framework of cooperative interfaces to complex artificial systems has recently received increasing attention. In this paper we present the IR-NLI II system, an expert interface that allows casual users to access online information retrieval systems and encompasses user modeling capabilities. More specifically, an illustration of the user modeling subsystem is given by describing the organization of the user model proposed for the particular application area, together with its use during system operation. The techniques utilized for the construction of the model are presented as well. They are based on the use of sterotypes, which are descriptions of typical classes of users. More specifically, they include both declarative and procedural knowledge for describing the features of the class to which the sterotype is related, for assigning a user to that class, and for acquiring and validating the necessary information during system operation. (INFORMATION PROCESSING & MANAGEMENT, Vol. 23, No. 4, pp. 305-320, 1987) A PROTOTYPE OF AN INTELLIGENT SYSTEM FOR INFORMATION RETRIEVAL: IOTA Y. Chiaramella and B. Defude, Laboratoire IMAG ``Genie Informatique,'' BP 68-38402 St. Martin d'Heres, France Recent results in artificial intelligence research are of prime interest in various fields of computer science; in particular we think information retrieval may benefit from significant advances in this approach. Expert systems seem to be valuable tools for components of information retrieval systems related to semantic inference. The query component is the one we consider in this paper. IOTA is the name of the resulting prototype presented here, which is our first step toward what we can an intelligent system for information retrieval. After explaining what we mean by this concept and presenting current studies in the field, the presentation of IOTA begins with the architecture problem, that is, how to put together a declarative component, such as an expert system, and a procedural component, such as an information retrieval system. Then we detail our proposed solution, which is based on a procedural expert system acting as the general scheduler of the entire query processing. The main steps of natural language query processing are then described according to the order in which they are processed, from the initial parsing of the query to the evaluation of the answer. The distinction between expert tasks and nonexpert tasks is emphasized. The paper ends with experimental results obtained from a technical corpus, and a conclusion about current and future developments. (INFORMATION PROCESSING & MANAGEMENT, Vol. 23, No. 4, pp. 285-303, 1987) TEXT SIGNATURES BY SUPERIMPOSED CODING OF LETTER TRIPLETS AND QUADRUPLETS Friedrich Gebhadt, Gesellschaft fur Mathematik und Datenverabeitung mbH, D- 5205 St Augustin, West Germany Text signatures are a condensed, coded form of a text; due to the reduced length, information is retrieved faster than with the full text if inverted files are not available. It has been proposed to base a particular form of signatures, the superimposed coding, on letter triplets (or quadruplets) rather than on complete words admitting in this way the masking of searchwords. This situation is analyzed here theoretically considering the unequal occurrence probabilities of the triplets; the results are compared with a set of experiments. It turns out that the signatures based on letter triplets produce too many false associations since the triplets occur in words other than the searchword. With quadruplets, the number of false associations might be tolerable. (INFORMATION SYSTEMS, Vol. 12, No. 2, pp. 151-156, 1987) CONCEPT RECOGNITION IN AN AUTOMATIC TEXT-PROCESSING SYSTEM FOR THE LIFE SCIENCES Natasha Vieduts-Stokolov, BIOSIS, 2100 Arch Street, Philadelphia, PA 19103 This article describes a natural-language text-processing system designed as an automatic aid to subject indexing at BIOSIS. The intellectual procedure the system should model is a deep indexing with a controlled vocabulary of biological concepts - Concept Headings (CHs). On the average, ten CHs are assigned to each article by BIOSIS indexers. The automatic procedure consists of two stages: (1) translation of natural-language biological titles into title-semantic representations which are in the constructed formalized language of Concept Primitives, and (2) translation of the latter representations into the language of CHs. The first stage is performed by matching the titles against the system's Semantic Vocabulary (SV). The SV currently contains approximately 15,000 biological natural-language terms and their translations in the language of Concept Primitives. For the ambiguous terms, the SV contains the algorithmical rules of term disambiguation, rules based on semantic analysis of the contexts. The second stage of the automatic procedure is performed by matching the title representations against the CH definitions, formulated as Boolean search strategies in the language of Concept Primitives. Three experiments performed with the system and their results are described. The most typical problems the system encounters, the problems of lexical and situational ambiguities, are discussed. The disambiguation techniques employed are described and demonstrated in many examples. (JOURNAL OF THE AMERICAN SOCIETY FOR INFORMATION SCIENCE, Vol. 38, No. 4, pp. 269-287, 1987) PROBABILISTIC RETRIEVAL AND COORDINATION LEVEL MATCHING Robert Losee, School of Library Science, University of North Carolina, Chapel Hill, NC 27514 Probabilistic models of document-retrieval systems incorporating sequential learning through relevance feedback may require frequent and time-consuming reevaluations of documents. Coordination level matching is shown to provide equivalent document rankings to binary models when term discrimination values are equal for all terms; this condition may be found, for example, in probabilistic systems with no feedback. A nearest-neighbor algorithm is presented which allows probabilistic sequential models consistent with two- Poisson or binary-independence assumptions to easily locate the ``best'' document using temporary sets of documents at a given coordination level. Conditions under which reranking is unnecessary are given. (JOURNAL OF THE AMERICAN SOCIETY FOR INFORMATION SCIENCE, Vol. 38, No. 4, pp. 239-244, 1987) OPTIMAL DETERMINATION OF USER-ORIENTED CLUSTERS Vijay V. Raghavan, The Center for Advanced Computer Studies, University of Southwestern Louisiana, Lafayette, LA 70504-4330 and Jitender S. Deogun, Department of Computer Science, University of Nebraska, Lincoln, NE 68588-0115 User-oriented clustering schemes enable the classification of documents based upon the user perception of the similarity between documents, rather than on some similarity function presumed by the designer to represent the user criteria. In this paper, an enhancement of such a clustering scheme is presented. This is accomplished by the formulation of the user-oriented clustering as a function-optimization problem. The problem formulated is termed the Boundary Selection Problem (BSP). Heuristic approaches to solve the BSP are proposed and some preliminary results that motivate the need for further evaluation of these approaches is provided. (PROCEEDINGS OF THE TENTH ANNUAL INTERNATIONAL ACM-SIGIR CONFERENCE ON RESEARCH & DEVELOPMENT IN INFORMATION RETRIEVAL, New Orleans, LA, USA, pp. 140-146, 1987) PROBABILISTIC SEARCH TERM WEIGHTING--SOME NEGATIVE RESULTS Norbert Fuhr and Peter Muller, TH Darmstadt, Fachbereich Informatik, 6100 Darmstadt, West Germany The effect of probabilistic search term weighting on the improvement of retrieval quality has been demonstrated in various experiments described in the literature. In this paper, we investigate the feasibility of this method for boolean retrieval with terms from a prescribed indexing vocabulary. This is a quite different test setting in comparison to other experiments where linear retrieval with free text terms was used. The experimental results show that in our case no improvement over a simple coordination match function can be achieved. On the other hand, models based on probabilistic indexing outperform the ranking procedures using search term weights. (PROCEEDINGS OF THE TENTH ANNUAL INTERNATIONAL ACM-SIGIR CONFERENCE ON RESEARCH & DEVELOPMENT IN INFORMATION RETRIEVAL, New Orleans, LA, USA, pp. 13-18, 1987) NON-HIERARCHIC DOCUMENT CLUSTERING USING THE ICL DISTRIBUTED ARRAY PROCESSOR Edie M. Rasmussen and Peter Willett, Department of Information Studies, University of Sheffield, Western Bank, Sheffield S10 2TN, U.K. This paper considers the suitability and efficiency of a highly parallel computer, the ICL Distributed Array Processor (DAP), for document clustering. Algorithms are described for the implementation of the single-pass and reallocation clustering methods on the DAP and on a conventional mainframe computer. These methods are used to classify the Cranfield, Vaswani and UKCIS document test collections. The results suggest that the parallel architecture of the DAP is not well suited to the variable-length records which characterize bibliographic data. (PROCEEDINGS OF THE TENTH ANNUAL INTERNATIONAL ACM-SIGIR CONFERENCE ON RESEARCH & DEVELOPMENT IN INFORMATION RETRIEVAL, New Orleans, LA, USA, pp. 132-139, 1987) QUALITY OF INDEXING IN ONLINE DATA BASES Howard D. White and Belver C. Griffith, College of Information Studies, Drexel University, Philadelphia, PA 19104 We describe practical tests by which the quality of subject indexing in online bibliographic data bases can be compared and judged. The tests are illustrated with 18 clusters of documents from the medical behavioral science literature and with terms drawn from MEDLINE, PsycINFO, BIOSIS, and Excerpta Medica. Each test involves obtaining a cluster of about five documents known on some grounds to be related in subject matter, and retrieving their descriptors from at least two data bases. We then tabulate the average number of descriptors applied to the documents, the number of descriptors applied to all and to a majority of the documents in the cluster, and the relative rarity of the applied descriptors. Comparable statistics emerge on how each data base links related documents and discriminates broadly and finely among documents. We also gain qualitative insights into the expressiveness and pertinence of the available indexing terms. (INFORMATION PROCESSING & MANAGEMENT, Vol. 23, No. 3, pp. 211-224, 1987) ------------------------------ END OF IRList Digest ********************