Information Retrieval List Digest 119 (July 6, 1992) URL = http://hegel.lib.ncsu.edu/stacks/serials/irld/irld-119 ========================================================================= Date: Mon, 6 Jul 1992 16:13:10 PST Reply-To: "Information Retrieval List" Sender: "Information Retrieval List" From: IRLIST Subject: IR-L Digest, Vol.IX,No.23, Issue 119 IRLIST Digest July 6, 1992 Volume IX, Number 23 Issue 119 ********************************************************** II. QUERIES B. Requests for Information 1. Fuzzy Software IV. PROJECT WORK C. Abstracts 1. IR-Related Dissertation Abstracts ********************************************************** II. QUERIES II.B.1. Fr: Diego Gachet Re: Fuzzy Software I'd like to know if anyone knows the existence of a fuzzy package accesible by a nonymous ftp. Thanks in advance. Diego ********************************************************** IV. PROJECT WORK Fr: Susanne M. Humphrey Re: Selected IR-Related Dissertation Abstracts The following are citations selected by title and abstract as being related to Information Retrieval (IR), resulting from a computer search, using BRS Information Technologies, of the Dissertation Abstracts Online database produced by University Microfilms International (UMI). Included are UMI order number, title, author, degree, year, institution; number of pages, one or more Dissertation Abstracts International (DAI) subject descriptors chosen by the author, and abstract. Unless otherwise specified, paper or microform copies of dissertations may be ordered from University Microfilms International, Dissertation Copies, Post Office Box 1764, Ann Arbor, MI 48106; telephone for U.S. (except Michigan, Hawaii, Alaska): 1-800-521-3042, for Canada: 1-800-268-6090. Price lists and other ordering and shipping information are in the introduction to the published DAI. An alternate source for copies is sometimes provided. Dissertation titles and abstracts contained here are published with permission of University Microfilms International, publishers of Dissertation Abstracts International (copyright by University Microfilms International), and may not be reproduced without their prior permission. AN University Microfilms Order Number ADGDX-92801. AU MANJI, KARIM A. TI PICTORIAL COMMUNICATION WITH COMPUTERS. IN Council for National Academic Awards (United Kingdom) Ph.D. 1990, 508 pages. SO DAI V52(03), SecB, pp1550. DE Computer Science. Artificial Intelligence. AB Available from UMI in association with The British Library. This thesis is concerned with an investigation of some of the ways in which pictorial interface systems might be used, within a workstation environment, to facilitate end-user communication with a computer. It commences with an overview of some of the background considerations that need to be taken into account in defining the scope of the work that has to be undertaken. Subsequently, a detailed description is given of the workstation environments and interaction techniques needed to support pictorial communication methods. The two major application areas underlying the work described in this thesis are: (1) electronic books and (2) expert systems. The use of pictorial forms to facilitate end-user interaction within each of these application areas is investigated. A variety of different types of interaction technique are considered and evaluated. A suitable underlying workstation architecture is fundamental to the construction of interactive computer systems to support this type of work. Within this thesis an appropriate architecture is identified. This architecture involves the use of a User Interface Management System (UIMS), a range of interaction methodologies and suitable data base support facilities. Examples of the ways in which these underlying technologies are used are given in various parts of the thesis. The major outcomes of this research have been: (1) the identification of appropriate hardware and software architectures to support the fabrication of electronic books; (2) the design and implementation of UIMS and data base facilities for the different electronic book systems that have been studied; (3) the design of end-user evaluation experiments; and (4) the provision of pictorial interface systems to a range of different expert systems. AN University Microfilms Order Number ADG91-22322. AU OWENS, CHRISTOPHER CHARLES. TI INDEXING AND RETRIEVING ABSTRACT PLANNING KNOWLEDGE. IN Yale University Ph.D. 1990, 216 pages. SO DAI V52(03), SecB, pp1551. DE Computer Science. Artificial Intelligence. AB Intelligent systems can reason about plan failures and other critical planning situations by retrieving and instantiating abstract knowledge structures that characterize those situations. This approach requires a vocabulary of thematic abstractions for representing and indexing critical planning situations, and a mechanism for recognizing situations to which the thematic abstractions apply. This dissertation presents a taxonomization of critical planning situations and an abstract indexing vocabulary, both derived from the study of human planning knowledge in the form of common advice-giving proverbs. It also present a parallel, incremental model of retrieval that interleaves and integrates the processes of abstract feature detection and memory search. Together, the vocabulary and the retrieval mechanism constitute a model of memory-based reasoning about the diagnosis and repair of plan failures. AN University Microfilms Order Number ADG91-24753. AU PEREZ-CARBALLO, JOSE F. TI DESIGN AND IMPLEMENTATION OF HYTEK: A KNOWLEDGE-BASED HYPERTEXT SYSTEM. IN New York University Ph.D. 1990, 172 pages. SO DAI V52(03), SecB, pp1552. DE Computer Science. Artificial Intelligence. Information Science. AB A Hypertext system is a text data base where the units of information are interlinked using pointers that the user can follow. We call the pointers "explicit links" (as opposed to computed or virtual links.) HyTeK provides a set of tools designed to help the user explore the information contained in the system. The information contained in the system is represented using at least one of the three following methods: fragments of full text, explicit links between fragments and a collection of frame-like objects organized in a taxonomy. Explicit links are used to represent discourse relationships between fragments of text. The frame-like objects, called "Topics", represent concepts in the domain of the text contained in the fragments. Topics are used to index the fragments for retrieval. The taxonomy of Topics represents some of the relationships between fragments that a traditional Hypertext System would represent using explicit links. HyTeK's query system uses the taxonomy of Topics in order to implement tools that allow the user to retrieve fragments selectively by their contents. A user queries the system by building a set of Topics in an interactive process of reformulation. Query reformulation is supported by a set of tools that allow the user to explore the space of Topics. The relationships between the Topics are used to define a similarity measure which is used to rank the target set of the query. This work describes an automatic indexing scheme, a query system and an extension of the Knowledge Representation (KR) system NIKL (KLONE) that was used in HyTeK to implement the taxonomy of Topics. A prototype of HyTeK was implemented in Common-Lisp in a Symbolics 3645 running Genera 7.2. The system has been extensively tested on several test collections of a total of 1000 fragments of text about AIDS treatments. The results indicate clear advantages over traditional Information Retrieval systems and suggest that the use of a KR system for the implementation of a query module for a Hypertext System is promising. AN University Microfilms Order Number ADG13-42840. AU QUBAIN, MAHER ISSA. TI A RELATIONAL DATABASE OF DIGITIZED IMAGES USING THE ADA PROGRAMMING LANGUAGE. IN California State University, Long Beach M.S. 1990, 188 pages. SO MAI V29(03) pp465. DE Computer Science. AB The goal of this thesis is to create an application, called CompuVision, that will be used as a developers tool which can store, retrieve and manipulate digitized video images and data related to those images in support of other application programs. CompuVision uses digitized pictures from a video camera, and provides facilities for an application programmer to develop applications using the visual database. The programmer will be allowed to create applications which associate pictures with relational database records. AN This item is not available from University Microfilms International ADGC1-82268. AU SHARP, BERNADETTE. TI ELABORATION AND TESTING OF NEW METHODOLOGIES FOR AUTOMATIC ABSTRACTING. IN Aston University (United Kingdom) Ph.D. 1989. SO DAI V52(03), SecC, pp461. DE Computer Science. AB The primary objective of this research was to understand what kinds of knowledge and skills people use in "extracting" relevant information from text and to assess the extent to which expert systems techniques could be applied to automate the process of abstracting. The approach adopted in this thesis is based on research in cognitive science, information science, psycholinguistics and textlinguistics. The study addressed the significance of domain knowledge and heuristic rules by developing an information extraction system, called INFORMEX. This system, which was implemented partly in SPITBOL, and partly in PROLOG, used a set of heuristic rules to analyse five scientific papers of expository type, to interpret the content in relation to the key abstract elements and to extract a set of sentences recognised as relevant for abstracting purposes. The analysis of these extracts revealed than an adequate abstract could be generated. Furthermore, INFORMEX showed that a rule base system was a suitable computational model to represent experts' knowledge and strategies. This computational technique provided the basis for a new approach to the modelling of cognition. It showed how experts tackle the task of abstracting by integrating formal knowledge as well as experiential learning. This thesis demonstrated that empirical and theoretical knowledge can be effectively combined in expert systems technology to provide a valuable starting approach to automatic abstracting. AN University Microfilms Order Number ADG91-16435. AU MOHAN, LIL. TI A FRAMEWORK FOR CONCEPTUAL MODELING AND GRAPHICAL INTERACTION WITH KNOWLEDGE-INTENSIVE DOMAINS. IN Purdue University Ph.D. 1990, 135 pages. SO DAI V52(03), SecB, pp1633. DE Engineering, Electronics and Electrical. Computer Science. Engineering, Industrial. Artificial Intelligence. AB The first part presents an interactive framework for modeling knowledge-intensive domains. We have designed and implemented a prototype framework SSONET (Structured Semantic Object NETwork) which has the potential to incorporate all the schema design and data management capabilities available in current Object-Oriented databases as well as the knowledge representation and reasoning capabilities available in artificial intelligence related systems. One goal of creating such a representational framework has been to impose a structural discipline on the creation and the manipulation of conceptual objects within Object-Oriented knowledge models. A fundamental design issue underlying the development of the SSONET modeling framework is the utilization of explicit links. The framework allows for the specification of meta-level conceptual relationships between the semantic links themselves. The second part presents a novel visual query language (VQL) for interacting with the SSONET knowledge model. VQL allows convenient access to the various types of knowledge captured by the knowledge model. It consists of a set of 'graphical primitives' along with a combination grammar for creating graphical queries. Apart from being able to create simple queries such as those that can be specified in SQL or QBE, VQL can be used for making queries on the ER model. Further, VQL handle complicated, indirect queries. Recursive queries on graph structures such as finding transitive closures of graphs can be pictorially specified in an elegant way. Perhaps the most powerful feature of VQL is that it provides high semantic expressibility (in being able to specify highly complex queries) without sacrificing ease of user query formulation. SSONET and VQL are embedded in an object-oriented visual knowledge base interaction environment that supports schema creation and manipulation as well as database querying and updation. The prototype framework has been implemented in Smalltalk-80 running on a Sun 3/60 workstation. The framework is illustrated through a particular domain knowledge model created for Computer Integrated Manufacturing applications. It is currently being used as a platform for designing various CIM applications such as knowledge based tools for engineering design, assembly planning and machine fault diagnosis. (Abstract shortened with permission of author.). AN University Microfilms Order Number ADG91-25400. AU SU, LOUISE T. TI AN INVESTIGATION TO FIND APPROPRIATE MEASURES FOR EVALUATING INTERACTIVE INFORMATION RETRIEVAL. IN Rutgers The State University of New Jersey - New Brunswick Ph.D. 1991, 307 pages. SO DAI V52(03), SecA, pp720. DE Information Science. Library Science. AB There is no clear guidance in IR literature about how to evaluate interactive information retrieval performance. Several criteria and measures have been proposed and used in evaluation. There is no agreement about what is a successful IR performance or which are the best existing evaluation measures. This study aims to identify appropriate measures for evaluating interactive IR performance, i.e. evaluative measures that can best reflect system's success in providing help for users' needs or problems. Twenty measures of IR performance were selected for study, representing four major evaluation criteria: relevance, utility, user satisfaction and efficiency. Four non-performance characteristics were also selected for subgroup analysis. User's judgment of system success in realistic IR situation was used as the devised criterion measure with which all other 20 measures were to be correlated. Six searchers and 40 end-users were recruited from Rutgers Online Automatic Retrieval Service. A computer search was conducted by a searcher in the normal manner, for each individual problem, in the user's presence and with his/her participation. The end-users were each responsible for the cost of his/her own search. This study showed that value of search results as a whole is the best single measure of interactive IR performance among the 20 measures selected. It is strongly correlated with other significant correlates of success: user's satisfaction with completeness of search results, and user's satisfaction with precision of the search. Precision, one of the most important traditional measures of effectiveness, was found to be not significantly correlated with success. Users appeared to be more concerned with absolute recall than with precision of the search. The study also identified two more basic factors which can account for much higher proportion of the total variance than value of search results as a whole alone can. The study suggested 17 new success categories for further investigation. Open issues and future research are also discussed. AN University Microfilms Order Number ADG13-42915. AU HENNAN, ROBERT JOHN. TI AN INTELLIGENT LIBRARIAN SYSTEM FOR ACCESSING AND LOADING LARGE DISTRIBUTED KNOWLEDGE DATABASES. IN University of Houston M.S. 1991, 178 pages. SO MAI V29(03) pp355. DE Library Science. Artificial Intelligence. Computer Science. AB Recent advances in computer, communications, and information storage technologies have provided opportunities to construct more sophisticated delivery systems for information services. These advances are crucial to information seekers because of the tremendous growth in the volume of online data that is threatening to overwhelm us. The problem is extracting useful information, i.e., knowledge, from the wealth of online information. Artificial Intelligence (AI) techniques will play a large role in addressing this problem. Some of the more advanced online systems are already using AI to a limited degree. Over the next several years, the role of AI will increase through necessity, since it holds the greatest promise for dealing with the information explosion. This thesis will present the most advanced techniques used by today's online systems and research prototypes. From these observations, a recommended composite system is proposed for an intelligent online knowledge database. AN University Microfilms Order Number ADG91-21316. AU BLUMENTHAL, KENT JULIUS. TI THE CONCEPTUAL STRUCTURE OF A THESAURUS FOR THE PARK, RECREATION AND LEISURE SERVICE PROFESSION. IN University of Maryland College Park Ph.D. 1990, 222 pages. SO DAI V52(03), SecA, pp1076. DE Recreation. Library Science. AB The purpose of this study was to develop the conceptual structure of a thesaurus for the park, recreation and leisure service profession, and to provide empirical evidence for the developed conceptual structure. A modified Delphi technique was employed as the research method. A series of three questionnaires were used for data collection. Panels were comprised of professional members from among six groups of the National Recreation and Park Association. Panelists rated 313 concepts according to their importance to the conceptual structure of a thesaurus, using a six point Likert-type scale. Principal Factor Analysis with Varimax Rotation identified 16 factors, based on 76 highly rated concepts that reached a high level of congruence among the panel. A one-way ANOVA found group type to be significant for the first (Leisure Behavior and Attitudes) and third (Environment and Natural Resources) factors. Leisure Behavior and Attitudes was significantly more important to educators and therapeutic recreation specialists, and Environment and Natural Resources was significantly more important for park resource professionals, when compared to other groups. It was concluded that 16 core dimensions comprise the conceptual structure of a thesaurus. These dimensions include concepts unique to specializations within the field, as well as concepts which characterize other disciplines. Theoretical and practical implications derived from results of this study pertain to the body of knowledge of the discipline, competency-based professional certification and accreditation guidelines, and usefulness of conceptually complete thesauri for automated information storage and retrieval systems. AN This item is not available from University Microfilms International ADGC1-79716. AU BIENNIER, FREDERIQUE. TT MODELLING HYPERDOCUMENT BASE AND A CONNECTIONIST METHOD FOR GUIDING BROWSING. TI MODELISATION D'UNE BASE D'HYPERDOCUMENTS ET METHODE CONNEXIONNISTE D'AIDE A LA NAVIGATION. LG FRENCH. IN Institut National des Sciences Appliquees de Lyon (France) DO. 1990, 256 pages. SO DAI V52(03), SecC, pp458. DE Computer Science. RC INSA DE LYON BIBLIOTHEQUE (LIBRARY), 20 AV. ALBERT EINSTEIN, 69621 VILLEURBANNE CEDEX, FRANCE. AB Hyperdocument bases must at least be as easy to use as paper documents. One of the readers' major problems is to select a path from a myriad of browsing possibilities along the defined links, adapted to their own goals in order to reach the information they need. First we propose a storage model for the hyperdocument base. By splitting the structure in three levels and an heavy use of persistent trees in each level, redundancy is avoided and several kinds of version can be stored. Then the documentary base is coupled to an associative epigenetic neural network. By running this network, according to particular activation rules, a path adapted to the users' defined needs is dynamically built. By this way, the system proposes the answers and their organizations which seem to best fit the users' needs. By using several simple parameters, the users can entirely control the system and adjust the answers to their particular needs by several refinements. AN This item is not available from University Microfilms International ADGC1-84418. AU STOGERER, JOSEF KARL. TT RETRIEVAL AND REPLACEMENT IN GRAPHICAL DATA. TI SUCHEN UND ERSETZEN IN BILDDATENBESTANDEN. LG GERMAN. IN Technische Universitaet Graz (Austria) Dr. 1991, 224 pages. SO DAI V52(03), SecC, pp461. DE Computer Science. RC TECHNISCHE UNIVERSITATSBIBLIOTHEK GRAZ, TECHNIKERSTRASSE 4, A-8010 GRAZ, AUSTRIA. AB Rapid distribution of graphics software in various application areas are accompanied by the storage of sometimes huge and mostly innumerable picture information "files". The casual user within such an environment simply has the problem of where to find "his" pictures. Therefore graphical retrieval and manipulation operations should be supported in a comfortable way by the basic (relational) data base management system. However, this is not the case in most graphics software systems today. Starting with a detailed theoretical discussion of graphical operators, two kernels of graphical retrieval and manipulation languages--PQL and SEPL--are presented. (1) Conceptually PQL--Picture Query Language--is based upon current query languages within relational data base systems. PQL was designed to be compatible as far as possible to the ISO standard SQL. However, PQL goes beyond SQL and includes graphical data types, support of general hierarchical structures, complex graphical predicates, recursive picture definition, similarity matches, etc. (2) SEPL--Set Oriented Picture Programming Language--is the kernel of a procedural graphics programming language. SEPL includes language elements for modelling and processing of complex objects and appropriate transitive operators. This is achieved by supporting orthogonal constructors for general purpose data types like sets, lists, multisets, etc. Furthermore, there are numerous predefined graphical data types according to PIC, a powerful and flexible picture storage format similar to CGM including concepts of PHIGS. Finally, data base aspects, e.g. ad hoc query, persistency, etc. are taken into account. These capabilities of PQL and SEPL represent a major step to reduce the semantic gap between current data base languages and conventional programming languages. ********************************************************** IRLIST Digest is distributed from the University of California, Division of Library Automation, 300 Lakeside Drive, Oakland, CA. 94612-3550. Send subscription requests to: LISTSERV@UCCVMA.BITNET Send submissions to IRLIST to: IR-L@UCCVMA.BITNET Editorial Staff: Clifford Lynch lynch@uccmvsa.ucop.edu or calur@uccmvsa.bitnet Nancy Gusack ncgur@uccmvsa.bitnet Mary Engle engle@cmsa.berkeley.edu or meeur@uccmvsa.bitnet The IRLIST Archives will be set up for anonymous FTP, and the address will be announced in future issues. To access back issues presently, send the message INDEX IR-L to LISTSERV@UCCVMA.BITNET. To get a specific issue listed in the Index, send the message GET IR-L LOG ***, where *** is the month and day on which the issue was mailed, to LISTSERV@UCCVMA.BITNET. These files are not to be sold or used for commercial purposes. Contact Nancy Gusack or Mary Engle for more information on IRLIST. The opinions expressed in IRLIST do not represent those of the editors or the University of California. Authors assume full responsibility for the contents of their submissions to IRLIST.