One should not throw out the baby with the bathwater, according to an old aphorism. Some popular recent positions in AI thinking have done just this, we suggest, by rejecting the useful idea of mental representations in their overenthusiastic zeal to correct some simplifications and naiveties in the way traditional AI ideas have sometimes been understood. These "situated" perspectives correctly emphasize that agents live in a social world, using their environments to help guide their actions without needing to always plan their futures in detail; but they incorrectly conclude that the very idea of mental representation is mistaken. This perspective has its intellectual roots in parts of recent sociological thinking which reject the entire fabric of western science. We discuss these ideas and disputes in the form of an illustrated fable concerning nannies and babies.

To assess the state of the art and identify issues requiring further investigation, a workshop on qualitative and abstract probability was held during the third week of November 1993. This workshop brought together a mix of active researchers from academia, industry, and government interested in the practical and theoretical impact of these abstractions on techniques, methods, and tools for solving complex AI tasks. The result was a set of specific recommendations on the most promising and important avenues for future research.

The vast amounts of on-line text now available have ledto renewed interest in information extraction (IE) systems thatanalyze unrestricted text, producing a structured representation ofselected information from the text. This paper presents a novel approachthat uses machine learning to acquire knowledge for some of the higher level IE processing. Wrap-Up is a trainable IE discourse component that makes intersentential inferences and identifies logicalrelations among information extracted from the text. Previous corpus-based approaches were limited to lower level processing such as part-of-speech tagging, lexical disambiguation, and dictionary construction. Wrap-Up is fully trainable, and not onlyautomatically decides what classifiers are needed, but even derives the featureset for each classifier automatically. Performance equals that of a partially trainable discourse module requiring manual customization for each domain.

This paper is a multidisciplinary review of empirical, statistical learning from a graphical model perspective. Well-known examples of graphical models include Bayesian networks, directed graphs representing a Markov chain, and undirected networks representing a Markov field. These graphical models are extended to model data analysis and empirical learning using the notation of plates. Graphical operations for simplifying and manipulating a problem are provided including decomposition, differentiation, andthe manipulation of probability models from the exponential family. Two standard algorithm schemas for learning are reviewed in a graphical framework: Gibbs sampling and the expectation maximizationalgorithm. Using these operations and schemas, some popular algorithms can be synthesized from their graphical specification. This includes versions of linear regression, techniques for feed-forward networks, and learning Gaussian and discrete Bayesian networks from data. The paper concludes by sketching some implications for data analysis and summarizing how some popular algorithms fall within the framework presented. The main original contributions here are the decompositiontechniques and the demonstration that graphical models provide a framework for understanding and developing complex learning algorithms.

This report summarizes this year's workshop and outlines WET to underwrite and support these workshops. Information Systems is also acknowledged. The Defense Advanced this year's workshop and outlines the philosophy behind this annual event. Computer-Supported Cooperative and present the best research Finally, I would like to thank V. Work gathering, which takes in that has a bearing on the "repersonalization Jagannathan for his great help and everyone from anthropologists to of computing," as Fernando expertise in workshop management futurists, this workshop focuses on flores, founder of Action Technologies, and Mary Carriger for relieving me of hardware and software that enables puts it.

This report stems from an April 1994 meeting, organized by AAAI at the suggestion of Steve Cross and Gio Wiederhold.1 The purpose of the meeting was to assist ARPA in defining an agenda for foundational AI research. Prior to the meeting, the fellows and officers of AAAI, as well as the report committee members, were asked to recommend areas in which major research thrusts could yield significant scientific gain -- with high potential impact on DOD applications -- over the next ten years. At the meeting, these suggestions and their relevance to current national needs and challenges in computing were discussed and debated. An initial draft of this report was circulated to the fellows and officers. The final report has benefited greatly from their comments and from textual revisions contributed by Joseph Halpern, Fernando Pereira, and Dana Nau.

Chevron Canada is decentralizing its computer environment from mainframes to PCs e Emission Reduction Research mounted infantry virtual environment) and Sun workstations. The DIVE belt provides the Chevron oil exploration crews will be N.J.) has developed the Batch Design ability to operate inside a virtual environment able to retrieve various types of well Kit, an expert system for optimizing without becoming tangled batch processes and minimizing pollution. Funded by the U.S. The system will help eliminate Army, the DIVE project is designed to ADVANTA Mortgage (San Diego, avoidable pollution and save pharmaceutical allow soldiers to operate within a virtual Cal.) has signed a license agreement and chemical manufacturers battlefield. VR is being used to demonstrate used as the focal point of exhibition fire engineering principles such The Santa Fe Institute (Santa Fe, stands designed by Photosound for as means of escape theory, fire modeling, N.M.) has won an ARPA grant of such pharmaceutical firms as Smith-human behavior, and spatial $323,000 for research on complex Kline Beecham. The system will be designed advanced computin arena.

What criteria should be used to select one semantic formalism over another? However, the scope of analyze and gain insight into (that is, models for circumscription, perfect convergence results linking aspects of not just model) such a task. Although much basic problems are NP hard (at best). Ginsberg and Hugh Holbrook work remains to be done, the consensus His point was that just confirming (Stanford University) showed seems to be that there is sufficient that this problem is indeed potentially that default reasoning could be used common ground to warrant serious nasty is not really surprising. Marco Cadoli and as well as to somehow cope with the significant computational advantages.

The Association for the Advancement of Artificial Intelligence (AAAI) held its 1994 Spring Symposium Series on 19-23 March at Stanford University, Stanford, California. This article contains summaries of 10 of the 11 symposia that were conducted: Applications of Computer Vision in Medical Image Processing; AI in Medicine: Interpreting Clinical Data; Believable Agents; Computational Organization Design; Decision-Theoretic Planning; Detecting and Resolving Errors in Manufacturing Systems; Goal-Driven Learning; Intelligent Multimedia, Multimodal Systems; Software Agents; and Toward Physical Interaction and Manipulation. Papers of most of the symposia are available as technical reports from AAAI.

Over 60 researchers from 10 countries took part in the Third Knowledge Discovery in Databases (KDD) Workshop, held during the Eleventh National Conference on Artificial Intelligence in Washington, D.C. A major trend evident at the workshop was the transition to applications in the core KDD area of discovery of relatively simple patterns in relational databases; the most successful applications are appearing in the areas of greatest need, where the databases are so large that manual analysis is impossible. Progress has been facilitated by the availability of commercial KDD tools for both generic discovery and domain-specific applications such as marketing. At the same time, progress has been slowed by problems such as lack of statistical rigor, overabundance of patterns, and poor integration. Besides applications, the main themes of this workshop were (1) the discovery of dependencies and models and (2) integrated and interactive KDD systems.