Information Retrieval List Digest 049 (February 15, 1991) URL = http://hegel.lib.ncsu.edu/stacks/serials/irld/irld-049 IRLIST Digest February 15, 1991 Volume VIII, Number 6 Issue 49 ********************************************************** I. NOTICES A. Meetings announcements/Calls for papers 1. 3rd Conference on Applied Natural Language Processing April 1-3, 1992 Trento, Italy 2. 14th International Conference on Computational Linguistics (COLING '92) July 23-28, 1992 Nantes, France IV. PROJECT WORK B. Bibliographies 1. Selected IR-related dissertation abstracts ********************************************************** I. NOTICES I.A.1. Fr: Donald Walker Re: 3rd Conference on Applied Natural Language Processing April 1-3, 1992 Trento, Italy Sponsored by Association for Computational Linguistics PURPOSE The focus of this conference is on the application of natural language processing techniques to real world problems. It will include invited and contributed papers, tutorials, an industrial exhibition, and demonstrations. A special video session is also being organised. The organizers want the conference to be as international as possible, and to feature the best applied natural language work presently available in the world. This conference follows on from those held in Santa Monica, California in 1983, and in Austin, Texas in 1988. AREAS OF INTEREST Original papers are being solicited in all areas of applied natural language processing, including but not limited to: dialog systems; integrated speech and natural language systems; machine translation; explanation and generation; database interface systems; tool development; text and message processing; grammar and style checking; corpus development; knowledge acquisition; lexicons; language teaching aids; evaluation; adaptive systems; multilanguage systems; multimedia systems; help systems; and other applications. Papers may discuss applications, evaluations, limitations, and general tools and techniques. Papers that critically evaluate a relevant formalism or processing strategy are especially welcome. REQUIREMENTS FOR SUBMISSION Authors should submit, by 10 September 1991, a) six copies of a full-length paper (min 9, max 18 double-spaced pages, minimum font size 12, exclusive of references); b) 16 copies of a 20-30 line abstract; c) a declaration that the paper has not been accepted nor is under review for a journal or other conference nor will it be submitted during the conference review period. Papers arriving after the deadline will be returned unopened. We regret that papers cannot be submitted electronically, or by fax. Papers should describe completed rather than intended work, identify distinctive aspects of the work, and clearly indicate the extent to which an implementation has been completed; vague or unsubstantiated claims will be given little weight. Both the paper and the abstract should include the title, the name(s) of the author(s), complete addresses and e-mail address. Papers from Europe and Asia should be sent to: Oliviero Stock (ANLP-3)phone: +39-461-814444 I.R.S.T. fax: +39-461-810851 38050 Povo (Trento), ITALYemail: stock@irst.it Papers from America and other continents should be sent to: Madeleine Bates (ANLP-3)phone: +1-617-8733634 BBN Systems & Technologies fax: +1-617-8733776 10 Moulton Streetemail: bates@bbn.com Cambridge, MA 02138, USA Authors will be notified of acceptance or rejection by 30 November 1991. Full-length versions of accepted papers, prepared according to instructions, must be received, along with a signed copyright release statement, by 15 January 1992. All papers will be reviewed by members of the program committee, which is co-chaired by Madeleine Bates (BBN Systems & Technologies) and Oliviero Stock (IRST) and also includes: Robert Amsler, MITRE Kathy McKeown, Columbia Univ. Giacomo Ferrari, Univ. of Pisa Sergei Nirenburg, Carnegie Mellon Univ. Eduard Hovy, USC/ISI Makoto Nagao, Kyoto Univ. Paul Jacobs, General Electric Remko Scha, Univ. of Amsterdam Martin Kay, Xerox PARC Karen Sparck Jones, Univ. of Cambridge Mark Liberman, Univ. of Pennsylvania Henry Thompson, Univ. of Edinburgh Paul Martin, MCC Wolfgang Wahlster, DFKI VIDEOTAPES Videotapes are sought that display interesting research on NLP applications to real-world problems, even if presented as promotional videos (not advertisements). An ongoing video presentation will be organized that will demonstrate the current level of usefulness of NLP tools and techniques. Authors should submit one copy of a videotape of at most 15 minutes duration, accompanied by a submission letter giving permission to copy the tape to a standard format and two copies of a one to two page abstract that includes: title, name and address and email or fax number of authors; tape format of the submitted tape (VHS, any of NTSC, PAL or SECAM); duration. The final tape format provided by the authors should be one of VHS, 75'' u-Matic, BVU, in any of NTSC, PAL or SECAM. Videotapes cannot be returned. Tape submissions should be sent to the same address as the papers (see above). The timetable for submissions, notification of acceptance or rejection, and receipt of final versions is the same as for the papers. See above for details. Tapes will be reviewed and selected for presentation during the conference. Abstracts of accepted videos will appear in the conference proceedings. We are also considering the possibility of producing a collection of video proceedings, for those videotapes that authors agree to distribute. A preliminary indication on this matter will be appreciated. DEMONSTRATIONS Beside demonstrations to be carried on within a regular booth at the industrial exhibition, there will be a program of demonstrations on standard equipment available at the conference (SUN's, MAC's, etc.). Anyone wishing to present a demo should send a one-page description of the demo and a specification of the system requirements by 1 December 1991 to Carlo Strapparavaphone: +39-461-814444 I.R.S.T. fax: +39-461-810851 38050 Povo (Trento), ITALYemail: strappa@irst.it PRIZE A prize will be given for the best nonindustrial demonstration. TUTORIALS The meeting will be preceded by one or two days of tutorials by noted contributors to the field. Responsible for tutorials: Jon Slackphone: +39-461-814444 I.R.S.T. fax: +39-461-810851 38050 Povo (Trento), ITALYemail: slack@irst.it WORKSHOPS Proposals for organizing workshops in Trento immediately after the conference can be addressed to Oliviero Stock at the above address. INDUSTRIAL EXHIBITION Facilities for exhibits will also be available. Persons wishing to arrange an exhibit should send a brief description together with a specification of physical requirements (space, power, telephone connections, table, etc.) by 1 September 1991 to Giampietro Carlevarophone: +39-461-814444 I.R.S.T. fax: +39-461-810851 38050 Povo (Trento), ITALYemail: carleva@irst.it GENERAL INFORMATION Local arrangements are being handled by Tullio Grazioli and Oliviero Stockphone: +39-461-814444 I.R.S.T. fax: +39-461-810851 38050 Povo (Trento), ITALYemail: interne@irst.it For information on the ACL, contact Donald E. Walker (ACL)phone: +1-201-8294312 Bellcore, MRE 2A379 fax: +1-201-4551931 445 South Street, Box 1910email: walker@flash.bellcore.com Morristown, NJ 07960, USA The conference is also supported by the European Coordinating Committee for Artificial Intelligence (ECCAI), the Italian Association for Artificial Intelligence (AI*IA) and Istituto Trentino di Cultura. ********** I.A.2. Fr: Donald Walker Re: 14th International Conference on Computational Linguistics (COLING '92) July 23-28, 1992 Nantes, France FIRST ANNOUNCEMENT AND CALL FOR PAPERS DATES: The conference will last five full days (not counting Sunday). Pre-COLING tutorials will take place on 20-22 July (2-1/2 days). ORGANIZERS: GETA and IMAG, Grenoble (F. Peccoud, Ch. Boitet, J. Courtin), Palais des Congres, Nantes (M. Gillet), Universite de Nantes (M.H. Jayez), EC2 (G. d'Aumale). PROGRAMME CHAIR: Prof. A. Zampolli, Universita di Pisa, ILC, via della Faggiola 32, I-56100 Pisa, ITALY (tel: +39.50.560481; fax: +39.50.589055). DEADLINES: Send six A4 or 8-1/2 by 11 inch copies of the full paper to Prof. Zampolli before 1 November 1991. Notifications of acceptance will be sent by 1 March 1992. Camera-ready copies of final papers conforming to the COLING-90 style sheet must reach GETA (GETA-IMAG, COLING-92, BP 53X, F-38041 Grenoble, FRANCE) by 1 May 1992. TOPICS: All topics in Computational Linguistics are acceptable. Papers concerning real applications will be especially welcome. A special session on language industry is planned. Please indicate main areas of papers using two-level categories: computational models and formalisms (in morphology, syntax, semantics, pragmatics, discourse, dialogue, ...), methods (symbolic, numerical, statistical, neural, ...), tools (specialized languages, environments), large-scale resources (textual, lexical, grammatical databases), applications (natural language interfaces, information retrieval, text generation, machine translation, machine aids to writing, translating, abstracting, learning, ...), hypermedia and natural language processing (integration of text, speech, graphics, video), generic questions in language industry (engineering, ergonomics, legal aspects, normalization, ...). TYPES OF PAPERS: Topical papers (maximum seven pages in final format) on crucial issues in Computational Linguistics, and project notes (maximum five pages). Only unpublished papers will be accepted. Papers should describe substantial and original work, especially new methodologies and applications. They should emphasize completed rather than intended work. PRELIMINARY SCHEDULE: Twelve 30-minute lecture slots daily (hopefully in only three parallel sessions) and three 30-minute demonstration slots during the lunch break (hopefully in at least ten parallel sessions). It should be possible to have lunch and go to two or even three demos. DEMONSTRATIONS: Demonstrations are strongly encouraged. A project note without a demo will have a lower probability of acceptance. With a demo, it will get three consecutive demo slots. A topical paper including a demo will be presented as a lecture and as a demo. LANGUAGES: One extra page will be allowed for a long abstract in English, if the paper is written in another language, or conversely (paper in English and long abstract in another language). Speakers not giving their talk in English are encouraged to use visual aids in English. EXHIBITION: An exhibition of language industry products will be organized in parallel by EC2, the well known organizer of the annual Avignon meetings on Expert Systems. Industrial firms are encouraged to present state-of-the-art NLP products. OTHER ACTIVITIES: A social programme will be proposed to participants and companions. Individual discovery is also possible, as Nantes and its region are culturally very active and full of picturesque places. Organized on behalf of the International Committee on Computational Linguistics Martin Kay, Palo Alto (President); Eva Hajicova, Prague (Vice President); Donald E. Walker, Morristown (Secretary General); Christian Boitet, Grenoble; Nicoletta Calzolari, Pisa; Brian Harris, Ottawa; David Hays, New York (Honorary); Kolbjorn Heggstad, Bergen; Hans Karlgren, Stockholm; Olga Kulagina, Moscow; Winfried Lenders, Bonn; Makato Nagao, Kyoto; Helmut Schnelle, Bochum; Petr Sgall, Prague; Yorick Wilks, Las Cruces; Antonio Zampolli, Pisa ********************************************************** IV. PROJECT WORK IV.B.1. Fr: Susanne 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 ADG90-14646. AU SMITH, ELIZABETH SZYSZKOWSKI. TI SEARCH STRATEGIES FOR COMPUTER-STORED INFORMATION: A MICROCOMPUTER TUTORIAL. IN University of South Florida Ph.D. 1989, 245 pages. DE Library Science. Education, Curriculum and Instruction. Education, Technology. AB This study developed and tested a microcomputer tutorial teaching end users basic search strategies for retrieving computer stored information, using Boolean Operators. The tutorial, based on Ausubel's Meaningful Learning Theory, Gagne's External Events of Instruction, and Novak's Concept Mapping Technique, consists of four Lessons and a Practice Exercise. Evaluation of the tutorial consisted of content validation, pilot testing, and criterion validation. Thirty undergraduate education majors with no prior knowledge of online searching were randomly assigned to either the control or the experimental group, each consisting of 15 students. The experimental group worked through the tutorial; the control group was given a control tutorial. After completing the programs, each member in both groups formulated three search statements on three topics selected by the researcher. Statements of both groups were input online into the ERIC database in natural language using Dialog. Searches were limited to 10 citations in full record format. The results yielded 638 documents, 234 from the control and 404 from the experimental group. Citations were evaluated for relevancy by the researcher using the binary system of judgment (yes, no). Relevancy was determined by descriptors or keywords in the abstract in correct context. Viability of the tutorial as a method of teaching search strategies was established: mean precision score for the control group was 6.1%, the experimental group 49.36%. An obtained t ratio of 8.72 was larger than the table value of 2.048 to.05 level of confidence. The mean precision of 49.36% compares favorably with precision figures in literature which range from 42% to 70%. Implications are manifold: information available on computer cannot be accessed by many. Thus the role of the librarian is changing from searcher of information to teaching searching skills, which can be enhanced with this tutorial. Similar role shifts for teachers can occur. Results indicate that persons with no knowledge of computer searching can learn searching skills quickly utilizing this tutorial. Judgments of relevancy, despite guidelines, are subjective. AN University Microfilms Order Number ADG90-30395. AU VAN FLEET, CONNIE JEAN. TI COMMUNICATION BETWEEN PUBLIC LIBRARIANS AND LIBRARY EDUCATORS AS REFLECTED IN THE PUBLIC LIBRARY JOURNAL LITERATURE: A REFERENCE ANALYSIS. IN Indiana University Ph.D. 1990, 179 pages. DE Library Science. Information Science. Education, Higher. AB Communication models of the sciences and professions indicate the presence of subcultures with varying role responsibilities which interact to provide a relevant knowledge base. A literature review of library and information science reveals parallel subcultures in public librarianship: public librarians and library educators. A reference analysis of public library literature suggests that these two groups interact by citing each others' work, but that the literature is dominated by the work of others, both as authors and as authors of referenced works. Additional analysis indicates that library educators tend to provide references with greater frequency, and that these tend to be explicit in form. This result is not unexpected, given the different approaches to literature and research necessitated by the different primary roles of educators and librarians. The hypotheses that library educators act as gatekeepers for the profession by providing references to a wider variety of disciplines, a wider variety of formats, and more recent publications are not supported, although various interpretations of these findings are suggested by examination of the literature. Further delineation of categories, analysis of reference function, and exploration of the distribution of dates of publication of referenced works may be important in revealing differences underlying superficial similarities in the referencing practices of the two groups. AN University Microfilms Order Number ADG89-16441. AU RAHIM, SAMSUDIN ABDUL. TI DETERMINANTS OF DIACHRONIC INFORMATION SEEKING IN AN ORGANIZATION: A MALAYSIAN CASE STUDY. IN The University of Wisconsin - Madison Ph.D. 1989, 277 pages. DE Mass Communications. Information Science. Psychology, Industrial. AB This study focused on diachronic information seeking in an organization. The objectives were (a) to propose a model of information seeking and assess its viability, (b) to determine the relationships between variables in predicting information seeking, and (c) to study the relative importance of different categories of variables in predicting information seeking. Respondents were Information Assistants from the Department of Information, Malaysia. They were measured on three types of information seeking related to their task assignment to disseminate information about race relations: information about the issue, information about the audience, and information from the organization. Data were collected by questionnaire and analyzed by multiple regression analysis. The findings generally support the three-stage model: recognition stage, situational assessment stage, and information seeking stage. The model helps to explain why personal and attitudinal variables have little main effect on information seeking. Their effects were mostly indirect, mediated by variables in the later stages of the information seeking model. Hypotheses were tested for the total sample and for two ethnic groups, Malays and non-Malays (Chinese and Indians). Of 23 hypotheses tested, only six were not supported in any of the analysis groups. Variables with main effect on information seeking include knowledge on issue, extrinsic motivation, intrinsic motivation, information adequacy, communication constraint, referent criterion, assessment of issue, involvement in issue, fatalism, role preparedness, job involvement, and training. Information seeking was much influenced by factors related to the issue and respondents' jobs. Communication factors were less important in determining information seeking. Availability of information helps to motivate seeking of information about the audience and from the organization. However, perception that lack of information would affect communication effectiveness was negatively associated with information seeking. Among the three types of information seeking, information about the audience was the least sought, reflecting the lack of emphasis given to the importance of incorporating information about the audience in communication planning. AN University Microfilms Order Number ADG90-27838. AU WYMA, RICHARD JAMES. TI INVOLVING CHILDREN AS ACTIVE AGENTS OF THEIR OWN TREATMENT: A META-ANALYSIS OF SELF-MANAGEMENT TRAINING. IN Fuller Theological Seminary, School of Psychology Ph.D. 1990 359 pages. DE Psychology, Clinical. AB A meta-analysis was performed on the published literature on self-management training interventions with children to test the following two hypotheses: (1) Self-management interventions which involve children to a greater extent in implementing their own treatment operations will obtain better outcomes than interventions in which children are less involved. (2) The magnitude of effect sizes obtained by different types of outcome measures will be accounted for, in part, by the level of generalization the outcome measure represented. A computer search of the PsychINFO database was performed for empirical studies investigating behavioral self-management, social skills training, cognitive-behavioral, and biofeedback and/or relaxation training to remediate clinically significant psychological, educational, or medical child problems. Ninety-one articles from 1968-1989 were identified. Both single-subject as well as between-groups experiments were included. Each study was coded on a variety of subject, treatment, and outcome measure variables. The "System for Describing Therapeutic Interventions" (Clement, 1980) provided a measure of percentage of child-implemented treatment operations. The Generalization Map (Drabman, Rosenbaum & Hammer, 1979) was used to identify the level of generalization for each outcome measure. Percentage of child-implemented treatment operations was not found to be positively correlated with mean treatment effect size. The correlations between percentage of child-implemented treatment operations and mean effect size for single-subject designs was -.31 (p $<$.0001) and.08 (n. s.) for between-groups designs. Level of generalization of the outcome measure obtained the predicted relationship with effect size, supporting the second hypothesis. Direct outcome measures yielded higher effect sizes than did outcome measures reflecting increasingly distant levels of generalization. For single-subject designs mean effect sizes for outcome measures reflecting 0, 1, 2, 3, and 4 dimensions of generalization were 3.20, 3.00, 2.68, 1.75, and 2.08, respectively. For between-groups designs mean effect sizes for outcome measures reflecting 0, 1, 2, and 3 dimensions of generalization were 1.34, 1.10, 0.61, and 0.49, respectively. The relationships of other subject, treatment, methodological, and outcome measure variables to treatment effect size were also analyzed. ********************************************************** 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@postgres.berkeley.edu calur@uccmvsa.bitnet Mary Engle engle@cmsa.berkeley.edu meeur@uccmvsa.bitnet Nancy Gusack ncgur@uccmvsa.bitnet The IRLIST Archives will be set up for anonymous FTP, and the address will be announced in future issues. These files are not to be sold or used for commercial purposes. Contact Mary Engle or Nancy Gusack 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.