D-Lib Magazine
September 1997

ISSN 1082-9873

Clips and Pointers

IJCAI-97 Workshop on AI in Digital Libraries
Moving from Chaos to (more) Order, August 24 1997, Nagoya, Japan

The IJCAI-97 Workshop on AI and Digital Libraries was held in Nagoya, Japan, on Sunday, August 24 1997, as part of the Workshop Program of the Fifteenth International Joint Conference on Artificial Intelligence (IJCAI). The purpose of the workshop, as outlined in the original Call for Papers, was to identify and highlight ways in which AI techniques could contribute to solving some of the challenges of building real world digital libraries, and was intended to build upon the successes of earlier IJCAI and AAAI (American Association for Artificial Intelligence) workshops on information access and navigation. Following on from the many useful AI techniques and approaches presented at these venues, the main aim of the workshop was to focus on how such techniques might be applied in the context of digital libraries.

For the purposes of the workshop, and borrowing from the CNI White Paper on Networked Information Discovery and Retrieval, a digital library was defined to be an electronic information access system offering users a coherent view of a selected, organized, and managed body of information. As interest increases in heterogeneous, information-rich repositories - arguably, proto-digital libraries - such as the Internet and intranets, a number of technical challenges will need to be met to satisfy individual users' demands. Key among these are information discovery and retrieval, user interface design, classification and indexing, content delivery and presentation, storage management and administration, scalability, and interoperability. Faced with such challenges, a number of international agencies (most noticeably NSF, NASA, and DARPA in the United States, and JIPDEC [Japanese Information Processing Development Center] in Japan) have recently made available substantial funding for digital library research. AI researchers are among those participating in such large-scale digital library initiatives.

Particular questions that the workshop participants were asked to address included: How easily do AI techniques for general information repositories, such as the WWW, migrate to the more structured domains of digital libraries? What other AI techniques (eg. knowledge representation, machine learning) can assist in improving access to and management of digital information? and What new challenges and opportunities do real world digital libraries bring to bear on AI research? Specific topics of interest that were listed in the Call for Papers included:

• Information discovery and retrieval:
  • information filtering; user modelling and adaptive profiling; use of user feedback; software agents for resource discovery; knowledge-based models of information retrieval.
• User interface design:
  • information access, browsing and navigation; visualization of large repositories; summarization and multiple perspective presentation of content; natural language processing.
• Classification and indexing:
  • information modelling; knowledge and metadata representation; automatic keyword extraction, classification and clustering; statistical, probabilistic, linguistic and other approaches to indexing; indexing of non-textual media such as image, sound, and video; distributed indexing.
• Architectural designs and issues:
  • interoperability; scaling and extensibility; cooperative information systems; information economies.

Nine reviewed papers plus one invited talk were accepted for presentation at the workshop. The invited talk was titled Next Generation Digital Library Program in Japan and was presented by Hiroshi Mukaiyama, a senior researcher at JIPDEC, the Japanese Information Processing Development Center. In his talk, Hiroshi described the organization and scope of the 5-year national digital library initiative in Japan. Details on this project can be found here. The following papers (Postscript versions of which can be downloaded) were presented:

• Preferential Models of Refinement Paths in Query by Navigation (Peter Bruza and B. van Linder), describing a method for integrating preference logic and information retrieval to promote information discovery.
• Natural Language Processing for Session-based Information Retrieval Interface on the Web (Geunbae Lee et al.), describing a session-based web IR system using natural language and discourse processing.
• Identifying and Tracking Changing Interests (Barry Crabtree and Stuart Soltysiak), describing work on deriving user interest profiles automatically by monitoring user web and email habits.
• Automated Query-relevant Summarization and Diversity-based Reranking (Jaime Carbonell et al.), describing a method of operationalizing a plausible definition of "best" relevant documents by combining relevance with novelty (non-redundancy).
• Service Classification in a Proto-organic Society of Agents (Peter Weinstein and William Birmingham), describing a multi-agent based approach to service classification in large, dynamic, distributed systems.
• Smart Documents, Mobile Queries: Information Provision and Retrieval using a Multi-agent System (Steve Marsh), describing a multi-agent community-based information sharing system to facilitate the distribution of and search for information across networks.
• Toward the Mediation of Semantic Gaps between Users and Digital Libraries (Masanori Sugimoto et al.), describing a system for assisting users in constructing their own personal digital libraries.
• Adding Intelligence to Digital Libraries (Floriana Esposito et al.), describing a prototypical digital library system that applies machine learning techniques to facilitate information capture and semantic indexing.
• Robustness and Scalability in Digital Libraries: A case for Agent-based Market Systems (Innes Ferguson et al.), describing a multi-agent based digital library framework for developing commercial-strength information and knowledge management tools.

All in all, the workshop was very productive and informative. There was a lot of interesting discussion concerning the various pieces of presented work, as well as a reasonable level of information and knowledge sharing among the people present. At this time, there are no concrete plans to follow up with a second workshop on AI in Digital Libraries (participants will be polled to establish the level of interest in this idea); however, arrangements have commenced to publish expanded versions of a selection of the workshop papers in Springer Verlag's International Journal on Digital Libraries (JODL). This special issue on AI in Digital Libraries will be guest edited by Innes Ferguson and Ed Durfee and will likely appear in early 1998. A Call for Papers for this issue will be published in the near future.

NSF-Funded Workshop on "Advances in Organizational and Social Informatics", November 9-11, 1997, Bloomington, Indiana

Rob Kling and Howard Rosenbaum of Indiana University are organizing a Workshop on "Advances in Organizational and Social Informatics" with National Science Foundation (NSF) sponsorship. The workshop, with 25 invited active researchers, will be held in Bloomington, Indiana (at Indiana University) on November 9-11, 1997. The main purposes of this workshop are to (1) identify the state of knowledge about Organizational and Social Informatics; (2) identify productive directions for new studies; and (3) help forge a research community.

Social Informatics is a relatively new term to refer to the bodies of knowledge about information technologies and social change that are variously found under labels such as "social impacts of computing," "social design," and "implementation studies" (see http://www.slis.indiana.edu/SI). The movement has emerged from a series of lively conversations that began in early 1996among scholars with an interest in advancing research about the social aspects of computerization. They were concerned that the diversity of labels for social studies of information technologies impeded disciplinary and cross-disciplinary communication about important research results and theories. Like the name "Human Computer Interaction", the name "Social Informatics" serves as a banner to organize people who conduct research in the field and as a pointer to a body of research that could help information professionals, such as those who are interested in having digital libraries (DLs) effectively support particular communities.

Organizational Informatics is a sub-area of Social Informatics. Findings and theories belong to Organizational Informatics when they can be characterized primarily in terms of the participants of specific organizations. Using this criterion, our understandings of the adoption, use, and impacts of groupware fall well within Organizational Informatics. In contrast, the Internet is used by millions of people outside of their work lives, and the character and consequences of the public's use of the Internet is a topic outside of Organizational Informatics, but within Social Informatics. Studies of the social scale use and social roles of digital libraries also fit within, or can be informed by, Social Informatics.

Some of the key Organizational and Social Informatics research areas include:

• impacts of information technologies on the structure and communication patterns of groups and organizations;
• social life with computer-mediated communication (CMC, CSCW, DLs, Electronic publishing);
• the social shaping of information systems;
• the production, distribution and consumption of electronic texts;
• the roles of information technologies in changing or reinforcing patterns of work life, community life, and the character of institutions.

There is a notable body of systematic research about each of these topics. But it is scattered across the journals and conferences of several disciplines including computer science, information science, information systems, and various social sciences.

For additional information about Social Informatics, see the Social Informatics home page, at: http://www.slis.indiana.edu/SI For additional information about the Workshop on "Advances in Organizational and Social Informatics" contact Professor Rob Kling:

Center for Social Informatics
School of Library and Information Science
10th and Jordan
Indiana University Bloomington, IN 47405-1801
http://php.ucs.indiana.edu/~kling
812-855-9763 Fax: 812-855-6166

The Art Museum Image Consortium: Licensing Museum Digital Documentation for Educational Use

J. Trant and D. Bearman, Archives & Museum Informatics[1]

On September 22, more than twenty of the largest art museums in the USA and Canada, including the two national galleries, will form the Art Museum Image Consortium (AMICO). An initiative of the Association of Art Museum Directors, AMICO will enable educational access to cultural heritage information by creating and licensing a collective digital library of images and documentation of works in museum collections. Its foundation reflects the commitment of museums to an emerging new educational opportunity.[2]

The founding of AMICO follows discussions earlier this year, that defined mechanisms for the creation of a multimedia digital library documenting and interpreting museum collections. The agreements that frame the operation of AMICO (see http://www.amn.org/AMICO) are articulated at the level of values and objectives and define the operation of the Consortium and the licenses it will offer.

Access to the documentation held in museum collections has long been a problem, but what was an irritation in traditional scholarship has become a bottleneck in the networked educational environment.[3] The digitized text, images, audio tours and multimedia being created in museums provide a rich source of content for on-line curriculum. Interest in the use of non-traditional sources is growing as technologies finally enable their reproduction and integration.[4] But the cultural heritage community has lacked the economic and distribution systems to support this use. Intellectual property frameworks have been in flux,[5] and Fair Use does not enable many desired educational applications. Museums have been without precedent or policy regarding new digital educational uses, and some have been reluctant to grant rights at all.[6]

A not-for profit Consortium that licenses the content of museums collectively is the most effective means of distributing digital information to the educational community. Membership enables museums to do a number of things they cannot do on their own, including:

• Creating a collective library of art for educational uses
• Reaching the educational community in a coordinated and cost effective way
• Leveraging influence with vendors to reduce costs of capturing data
• Enabling access to collective funding to pursue educational goals
• Assisting members to improve their information infrastructures and documentation practices
• Negotiating digital rights with artists, artists estates and museums in other countries
• Providing members access to each others holdings for educational use

Collaborating in AMICO reduces risks for individual museums through collective decision-making, common standards and guidelines and shared expertise. Income supports services which enable members to contribute to the shared Library. These may include:

1) Technology Information: "best practice" guidelines, "frequently asked questions", standards for data capture, advice on hardware and software, application guidelines, training, research and liaison with standards developers.

2) Data Enhancement: data value standardization, unique identifiers, watermarking of images, subject indexing, metadata augmentation, thesaural explosion of terms in controlled vocabularies, markup of text to SGML, and mapping institutional data to export standards.

3) Catalog Management: creating an integrated, publicly accessible directory with many access points, so educators can identify works they have licensed and may use through AMICO, and the public can seek further rights (including commercial use) from individual museum members.

4) Rights Management: defining rights management data requirements, creating searchable rights systems, negotiating with individual rights holders and their collectives, writing model licensing agreements, developing terms of licenses for schools, school districts and public libraries, and drafting and disseminating user training materials.

5) Customer Services: monitoring and analyzing uses and users, identifying users needs, and promoting innovative educational uses of museum digital content.

6) Collaborative Partnering: developing liaisons with technology firms, funding sources, standards organizations, telecommunication providers, and others.

As a museum-owned collective, AMICO will be well positioned to collaborate with government sponsored agencies in Europe and elsewhere, following the example of sectors such as music.

Rather than fearing the simulacra of digital reproduction, museums are exploring the potential of digital technologies for developing increased familiarity with, and knowledge of the world's material culture.[7] In an era of shrinking budgets and challenges to the very existence of some national arts institutions, museums can make cultural resources relevant and accessible to educators and students. AMICO is the first step towards the creative use of these resources to illustrate and explore the Humanities. Its members are banking on the development of a visually and culturally literate public that feels an ownership of museum collections, as a consequence of having access to digital documentation in the course of their education.

Notes

[1]Archives & Museum Informatics has been retained by the AAMD to facilitate the planning and organisation of AMICO.

[2]Participants also recognized that by creating this educational resource they might gain higher visibility for digital content that individual museums could themselves license for single-item commercial uses.

[3]At a meeting sponsored by the Getty Art History Information Program in March 1994, leaders in the educational community identified three major barriers to the integration of visual information into cultural heritage networks, all of which are still present: a critical mass of digital content available for educational use, effective intellectual property rights clearance mechanisms, and adequate standards. See Getty AHIP, Initiative on Electronic Imaging & Information Standards, Imaging Initiative Meeting Report (March 3-4, 1994).

[4]Clifford Lynch, "The uncertain Future for Digital Visual Collections in the University", Archives and Museum Informatics, a cultural heritage informatics quarterly, Vol. 11, no 1, 1997, 5-13.

[5]The issues in moving from experiment to reality are explored by David Bearman in "New Models for Administering Cultural Intellectual Property", The Wired Museum, Washington, DC: American Association of Museums, 1997, 231-266.

[6] These issues are explored in detail in David Bearman and Jennifer Trant ÒMuseums and Intellectual Property: Rethinking Rights Management for a Digital WorldÓ, in Visual Resources, Vol. XII, no. 3-4, pp. 269-280. Licensing images, and the inter-related and layered rights questions, are outlined in J. Trant, "Exploring New Models for Administering Intellectual Property: The Museum Educational Site Licensing Project", Digital Imaging Access and Retrieval, Papers presented at the 1996 Clinic on Library Applications of Data Processing, March 24-26, 1997, Graduate School of Library and Information Science, University of Illinois at Urbana-Champagn, ed. by P. Bryan Heidorn and Beth Sandore, University of Illinois, 1997.

[7]Maxwell L. Anderson, "Introduction", The Wired Museum, 1997, pp. 11-34.

In Print

• Special Issue of Serials Librarian to be Devoted to Electronic Journals

  Wayne Jones
  Head of Serials Cataloging, MIT Libraries
  Cambridge, Massachusetts

  A new book about electronic serials is scheduled to be published by Haworth Press next spring. The book, a special issue of Serials Librarian, includes articles on a wide range of topics, including: publishing, pricing, copyright, acquisitions and collection development, cataloging and metadata, preservation and archiving, descriptions of various local/national/international projects, indexing, uniform resource identifiers, and citation. For more information, including abstracts of the articles, see the website at <http://web.mit.edu/waynej/www/onlineserials.htm>, or contact the editor, Wayne Jones, at [email protected]. The full-text of a few of the articles will also be posted to the site over the next months.

• Metadata Summit: Meeting Report, Willy Cromwell-Kessler and Ricky Erway.
  This report summarizes an invitational "summit meeting" on metadata issues organized by the Research Libraries Group on July 1, 1997. The "summit" proposed to identify metadata issues with special relevance for the research community in order to formulate, if possible, a common action agenda. Discussion was centered on the potential role of the Dublin Core elements as a mechanism for discovery of information resources available via the Internet, but which are not static HRML-Web pages.

• Network Wizards, Menlo Park, California, has released its semi-annual survey results (July 1997) on the size of the Internet. The report includes various tabular summaries based on top-level domain and host names as well as the authors' observations.

• The final report from the 1996 conference on research and development in electronic records, sponsored by the University of Michigan, School of Information and the Bentley Historical Library, has been posted together with the conference agenda. The report addresses issues in archiving, preservation, and related technologies.

• The Regional Alliance for Preservation is a one-year project involving four participants: the Upper Midwest Conservation Association, National Endowment for the Humanities, AMIGOS Preservation Service, Northeast Document Conservation Center, and Southeastern Library Network Preservation Services. The Alliance is testing two methods for collaborating: sharing training materials through a newsletter, and sharing equivalent information via a web site.

Point-to-Point

• The Canadian Initiative on Digital Libraries (CIDL), which represents a new alliance of Canadian libraries interested in improving communication and coordination in the development of Canadian digital library resources has been organized. It will provide a forum for:
  • sharing information and experience,
  • discussing best practices in areas such as digitization, metadata, encoding, rights management, etc.,
  • defining roles and responsibilities for long-term archiving,
  • raising awareness nationally and internationally,
  • and discussing many other issues including training, funding, interoperability, public access, etc.
  The bilingual, French-English site contains useful pointers to resources in digital libraries technologies as well as news on activities of the Initiative.

Goings On

• The Fall 1997 meeting of the Association of Information and Dissemination Centers (ASIDIC) will meet in Seattle, Washington, October 21-23, 1997. The program covers a number of issues relating to information management including databases, metadata, search, and simulation technologies.

• Beyond Print: Scholarly Publishing and Communication in the Electronic Environment will be held at the University of Toronto at Scarborough, Canada, September 26-27, 1997. The conference will deal with a number of issues related to the economics of publishing, peer review and quality control, the role of government agencies and university administrators, accessibility, intellectual property, and retrieval, archiving, and technology.

• Copyright Law in the Digital World: Fair Use, Education and Libraries after CONFU (The Conference on Fair Use) is a "town meeting" to be held on the campus of Reed College in Portland, Oregon, September 27, 1997. This is the fourth in a series of programs supported by the Samuel H. Kress Foundation to help educators, librarians, and other interested parties learn more about the CONFU process, the proposed guidelines, and fair use.

• The Fall 1997 Symposium for the American Medical Informatics Association (AMIA) will meet in Nashville, Tennessee, October 25-29, 1997. The themes of the conference are "The Emergence of 'Internetable' Health Care" and "Systems That Really Work!" Individual sessions are devoted to Java implementation, clinical information systems, data mining and software agents, and ethical challenges.

• Information Seeking in Context: an International Conference on Information Needs, Seeking and Use in Different Contexts will be held August 13-15, 1998 in Sheffield, UK. The Call for Proposals closes October 31, 1997. Details on submission as well as information on the program, registration, and accommodations are available through the web site.

• The Text Encoding Initiative: Tenth Anniversary User Conference will be held at Brown University in Providence, Rhode Island, from November 14 -to November 16, 1997. Although the event celebrates 10 years of work within the TEI, this is the first users conference that has been organized. Topics to be covered include but are not limited to the following:
  • use of the TEI scheme in a particular project or in a particular application area or discipline
  • interests of various user communities (e.g., builders and designers of electronic text centers, digital libraries, language corpora, electronic editions, multi-media databases, etc.)
  • evaluation in the context of a particular class of texts or in a given type of scholarly research
  • encoding problems and solutions that have arisen or may arise, such as unusual or complex text types, multi-media, multiple views or information types, multi-lingual data and internationalization, textual variation, overlap, etc.
  • customization and extension of the TEI for particular application areas and text types
  • experiences based on using off-the-shelf software with TEI documents, or developing software to handle TEI material
  • markup theory and markup architectures

Pointers in This Column:

1996 Electronic Records Conference Report University of Michigan, School of Information and The Bentley Historical Library	http://www.si.umich.edu/e-recs/
1997 Survey Results Network Wizards	http://www.nw.com/zone/WWW/report.html,
Adding Intelligence to Digital Libraries Floriana Esposito et al.	http://www.dlib.com/people/innes/aiindl/papers/esposito.ps
American Medical Informatics Association (AMIA) Fall 1997 Symposium October 25-29, 1997 Nashville, Tennessee	http://www.amia.org/f97flyer.html
Art Museum Image Consortium (AMICO)	http://www.amn.org/AMICO
Association of Information and Dissemination Centers (ASIDIC) Fall 1997 Meeting October 21-23, 1997 Seattle, Washington	http://www.accessinn.com/seattl~1.htm
Automated Query-relevant Summarization and Diversity-based Reranking Jaime Carbonell	http://www.dlib.com/people/innes/aiindl/papers/carbonell.ps
Beyond Print: Scholarly Publishing and Communication in the Electronic Environment September 26-27, 1997 University of Toronto at Scarborough, Canada	http://citd.scar.utoronto.ca/EPub/agenda.html
Canadian Initiative on Digital Libraries	http://www.nlc-bnc.ca/cidl/
Coalition for Networked Information (CNI) White Paper on Networked Information Discovery and Retrieval	http://www.cni.org/projects/nidr/
Copyright Law in the Digital World: Fair Use, Education and Libraries after CONFU (The Conference on Fair Use) "Town Meeting" Reed College, Portland, Oregon September 27, 1997	http://oregon.uoregon.edu/~csundt/pdx.htm
E-Serials (Serials Librarian)	http://web.mit.edu/waynej/www/onlineserials.htm
Identifying and Tracking Changing Interests Barry Crabtree and Stuart Soltysiak	http://www.dlib.com/people/innes/aiindl/papers/crabtree.ps
Information Seeking in Context: An International Conference on Information Needs, Seeking and Use in Different Contexts August 13-15, 1998 Sheffield, UK	http://panizzi.shef.ac.uk/david_allen/ISIC/isic.html
International Conference on Artificial Intelligence (IJCAI) Call for Papers	http://www.dlib.com/people/innes/aiindl/cfp.html
International Journal on Digital Libraries (JODL).	http://cimic.rutgers.edu/jodl/
Japanese Information Processing Development Center (JIPDEC)	http://www.gip.jipdec.or.jp/english/project-e/project27-e.html
Metadata Summit: Meeting Report Research Libraries Group	http://www.rlg.org/meta9707.html,
Natural Language Processing for Session-based Information Retrieval Interface on the Web Geunbae Lee et al.	http://www.dlib.com/people/innes/aiindl/papers/lee.ps
Preferential Models of Refinement Paths in Query by Navigation Peter Bruza and B. van Linder	http://www.dlib.com/people/innes/aiindl/papers/bruza.ps
Regional Alliance for Preservation	http://www-cpa.stanford.edu/cpa/misc/rap/
Robustness and Scalability in Digital Libraries: A Case for Agent-based Market Systems Innes Ferguson et al.	http://www.dlib.com/people/innes/aiindl/papers/ferguson.ps
Service Classification in a Proto-organic Society of Agents Peter Weinstein and William Birmingham	http://www.dlib.com/people/innes/aiindl/papers/weinstein.ps
Smart Documents, Mobile Queries: Information Provision and Retrieval using a Multi-agent System Steve Marsh	http://www.dlib.com/people/innes/aiindl/papers/marsh.ps
Social Informatics Home Page	http://www.slis.indiana.edu/SI
Text Encoding Initiative: Tenth Anniversary User Conference November 14-16, 1997 Brown University, Providence, Rhode Island	http://www.stg.brown.edu/webs/tei10/index.html
Toward the Mediation of Semantic Gaps between Users and Digital Libraries Masanori Sugimoto et al.	http://www.dlib.com/people/innes/aiindl/papers/sugimoto.ps">
USA Digital Libraries Initiative (NSF, NASA, and DARPA)	http://www.cise.nsf.gov/iris/DLHome.html

Copyright © 1997 Corporation for National Research Initiatives

hdl:cnri.dlib/september97-clips