Abstract: We propose constraints as the appropriate computational constructs for the design of agents with the task of selecting, merging and managing electronic information coming from such services as Internet access, digital libraries, E-mail, or on-line information repositories. Speci cally, weintroduce the framework of Constraint-Based Knowledge Brokers, which are concurrent agents that use so-called signed feature constraints to represent partially speci ed information and can exibly cooperate in the management of distributed knowledge. We illustrate our approach by several examples, and we de ne application scenarios based on related technology such asTelescript and work ow management systems. Key Words: multiagent coordination, agent-interaction, distributed problem solving, signed feature constraints, negotiation, cooperation strategies.

Access to on-line information via the Web is exploding. Index and retrieval engines already start to integrate a huge variety of heterogeneous repositories. However, the heterogeneity issue remains, both in terms of the search formats and the formats of the result pages. In this paper,

This paper describes WebFlow, an environment that supports distributed coordination services on the World Wide Web. WebFlow leverages the HTTP Web transport protocol and consists of a number of tools for the development of applications that require the coordination of multiple, distributed servers. Typical applications of WebFlow include distributed document workspaces, inter/intraenterprise workflow, and electronic commerce. In this paper we describe the general WebFlow architecture for distributed coordination, and then focus on the environment for distributed workflow.

The Internet and the World-Wide Web (WWW) are revolutionizing knowledge exchange by linking heterogeneous information repositories into a kind of gigantic world-wide digital library. Yet up until now, knowledge management on the WWW has mainly been provided by navigation tools like Mosaic and Netscape, and by engines like Alta Vista, Lycos and Yahoo which support navigation by automating the search for user-relevant WWW sites. The simplicity of this paradigm has been the key to the initial success of the Web infrastructure but now falls short of more complex applications needed by an ever-growing community of users. Prominent among these needs is flexible information gathering from multiple knowledge sources to ad-hocratically serve the requests of specific user groups. For instance, Network Publication Systems (NPS) for large organizations need flexible integration of enquiry information like Who's Who services and tables of contents of journals with E-print archival material, as well as flexible adaptation of local query services. Agent technology can provide the right answer to these demands. In this paper, we describe agent-based information gathering on the WWW in the context of a NPS for the European Physicist Society. In our approach, we exploit constraints to implement information gathering with maximal flexibility.

t be fully satisfied may still obtain results in the form of partial objects that refine the initial requests by instantiating some of its attributes or by adding new attributes. Efficient constraint solvers allow for local filtering/sifting of information. CBKB-related projects include: COIN (MIT), Harvest (Univ. Colorado), Garlic (IBM Almaden), HERMES (Univ. Maryland), Information Manifold (AT&T), InfoSleuth (MCC), KRAFT (Univ. Aberdeen / Wales / Liverpool, BT), SIMS (ISI), and TSIMMIS (Stanford Univ.). Besides information retrieval from the World Wide Web, another possible application of the CBKBmodel is dynamic document composition from on-line repositories: the widespread availability of new electronic sources of information such as E-mail, the Web, and other on-line information repositories also multiplies the number of electronic documents available. Documents can now be built dynamically by accessing and combining information existing over distributed sources. Hierarchical

Abstract: With the development and di usion of the Internet worldwide connection, a large amount of information is available to the users. Methods of information ltering and fetching are then required. This paper presents two approaches. The rst concerns the information ltering system ProFile based on an adaptation of the generalized probabilistic model of information retrieval. ProFile lters the netnews and uses a scale of 11 prede ned values of relevance. ProFile allows the user to update on{line the pro le and to check the discrepancy between the assessment and the prediction of relevance of the system. The second concerns ABIS, an intelligent agent for supporting users in ltering data from distributed and heterogeneous archives and repositories. ABIS minimizes user's e ort in selecting the huge amount ofavailable documents. The ltering engine memorizes both user preferences and past situations. ABIS compares documents with the past situations and nds the similarity scores on the basis of a memory-based reasoning approach.

The combination of push and pull information retrieval technology can benefit from collaborative information gathering schemes. In this paper we outline an approach that combines a pull technology (Knowl- edge Brokers) with a push technology (Knowledge Pump) in a community-based environment. Appropriate toolkits supporting synchronous collaboration on the World Wide Web are briefly introduced. A PlaceWare-based implementation illustrates the synchronous construction of queries and the publication of search activities and results supported by the push component.

Intelligent brokering of information needs algorithmically tractable knowledge representations and corresponding constraint solvers to tackle the satisfiability problems involved. Feature constraints emerged soon as a convenient and elegant choice of knowledge representation. The full fragment of feature constraints, where sorts and features are allowed to be combined by all the logical connectives (conjunction, disjunction, negation and quantifiers), although very expressive, is hardly tractable. On the other hand, the subfragment called "basic feature constraints" (BFC), where negation and disjunction are simply forbidden, unneccesarily restricts applications such as knowledge brokers in an intolerable way. The main contribution of this paper is twofold. On the one hand, it presents a fragment of feature constraints, called "signed feature constraints" (SFC), which allows limited use of negation, precisely capable of expressing the kind of operations needed during intelligen...

Today in physics, more and more scientific documents are prepared electronically, are posted and archived. At the same time, an increasing number of heterogeneous Web-servers are dedicated to hosting such archives. While this trend makes available more relevant scientific information, it also complicates the task of finding, retrieving, processing and printing scientific documents. As a consequence physicists could benefit from an efficient user-friendly tool that integrates functionalities for document retrieval and distribution. In this paper, we present a solution to this problem. Our approach links two systems together: an agent-based search engine for document retrieval and a printing-on-demand service that handles printing, binding and shipping of retrieved documents to readers who request them. We discuss a prototype already available that enables document retrieval from a number of Web servers, including the global network of physics preprints, set up by P. Ginsparg, and the PhysDoc broker which integrates 1000 distributed European Web-servers maintained by local physics departments.

ea: History When the Centre was founded in 1993 a natural language group was created first around a core team of existing Xerox researchers. Then a second group, originally called "Distributed Systems," was initiated to focus on distributed software systems and applications. The distributed systems group was started with people whose main competencies were the theoretical foundations of distributed computing. The group's initial focus was strongly application-oriented and its first goals were to identify application domains of interest to the corporation and sufficiently novel to justify a significant research investment. At the time, the corporation was undergoing a major shift, the result of which was to stress the importance of software in addition to traditional products that had made the corporation successful (copiers, printers etc.). At the same time, the extraordinary development of the Internet led to the notion that software components, even those embedded inside sta