Department of Computer Science, Carnegie Mellon University 5000 Forbes Ave, Pittsburgh, PA 15213 Abstract Dynamic Programming, Q-Iearning and other discrete Markov Decision Process solvers can be -applied to continuous d-dimensional state-spaces by quantizing the state space into an array of boxes. This is often problematic above two dimensions: a coarse quantization can lead to poor policies, and fine quantization is too expensive. Possible solutions are variable-resolution discretization, or function approximation by neural nets. A third option, which has been little studied in the reinforcement learning literature, is interpolation on a coarse grid. In this paper we study interpolation techniques thatcan result in vast improvements in the online behavior of the resulting control systems: multilinear interpolation, and an interpolation algorithm based on an interesting regular triangulation of d-dimensional space.

New additions in 1998 (UAI-98), July 24-26, 1998 February 1, 1998, use of the new 650 will be the Integrated Systems Track of --papers due February 23, 1998 code will be mandatory; so, please update the technical program and the Intelligent --www.uai98.cbmi.upmc.edu The Computing Research Association All regular members in good February 27: Intelligent Systems is planning a Career Development standing are encouraged to consider Demonstration proposals due. Details can be found at www. (at least one from a current is available by writing to ncai@ cra.org. AAAI fellow) must accompany nominations. Nomination materials are also available on AAAI is delighted to announce the AAAI-98 Workshop Program.

This article posits the idea of robot learning as a new subfield. The results of the Robolearn-96 Workshop provide evidence that learning in modern robotics is distinct from traditional machine learning. The article examines the role of robotics in the social and natural sciences and the potential impact of learning on robotics, generating both a continuum of research issues and a description of the divergent terminology, target domains, and standards of proof associated with robot learning. The article argues that although robot learning is a new subfield, there is significant potential for synergy with traditional machine learning if the differences in research cultures can be overcome.

Machine-learning research has been making great progress in many directions. This article summarizes four of these directions and discusses some current open problems. The four directions are (1) the improvement of classification accuracy by learning ensembles of classifiers, (2) methods for scaling up supervised learning algorithms, (3) reinforcement learning, and (4) the learning of complex stochastic models.

In recent years, there has been a resurgence in research on empirical methods in natural language processing. These methods employ learning techniques to automatically extract linguistic knowledge from natural language corpora rather than require the system developer to manually encode the requisite knowledge. The current special issue reviews recent research in empirical methods in speech recognition, syntactic parsing, semantic processing, information extraction, and machine translation. This article presents an introduction to the series of specialized articles on these topics and attempts to describe and explain the growing interest in using learning methods to aid the development of natural language processing systems.

Autonomous agents are computer systems that are capable of independent action in dynamic, unpredictable environments. Agents are also one of the most important and exciting areas of research and development in computer science today. Agents are currently being applied in domains as diverse as computer games and interactive cinema, information retrieval and filtering, user interface design, and industrial process control. Agents '98 will build on the enormous success of the First International Conference on Autonomous Agents (Agents '97), held in Marina del Rey in February 1997. The conference welcomes submissions of original, high quality papers and videos with summaries concerning autonomous agents in a variety of embodiments and playing a variety of roles in their environments.

Does machine learning really work? Yes. Over the past decade, machine learning has evolved from a field of laboratory demonstrations to a field of significant commercial value. Machine-learning algorithms have now learned to detect credit card fraud by mining data on past transactions, learned to steer vehicles driving autonomously on public highways at 70 miles an hour, and learned the reading interests of many individuals to assemble personally customized electronic newsAbstracts. A new computational theory of learning is beginning to shed light on fundamental issues, such as the trade-off among the number of training examples available, the number of hypotheses considered, and the likely accuracy of the learned hypothesis. Newer research is beginning to explore issues such as long-term learning of new representations, the integration of Bayesian inference and induction, and life-long cumulative learning. This article, based on the keynote talk presented at the Thirteenth National Conference on Artificial Intelligence, samples a number of recent accomplishments in machine learning and looks at where the field might be headed. [Copyright restrictions preclude electronic publication of this article.]

The Fifth International Conference on User Modeling (UM-96) is part of a recently established, biennial conference series that provides a forum for researchers in the field of user modeling and user-adapted interaction. The next major software revolution after graphic user interfaces will be software that adapts itself to the user. By adapting to the user's needs, preferences, knowledge, language, and even moods, software will attain new levels of usability and broad acceptance that would not be possible without built-in models of the user. This conference series provides a forum for recent research in the field, ranging from theoretical foundations to implemented systems to controlled studies of the human-computer interfaces of user-adapted systems.

Interestingly, the advantage for more effective access to traditional applications of agent of modeling others is diminished global and corporate information modeling, which requires an agent to when the volatility of the domain is repositories. These repositories are model the problem-solving processes low. Thus, the models of other agents increasingly multimedia, including of the interacting human to provide are more useful in variable domains, text, audio, graphics, imagery, and video. Now attention has turned appropriate feedback. Ole Mengshoel when they are a particularly valuable guide to predict what the other rational toward the problem of processing and D. C. Wilkins's (both of University agents are going to do. and managing multiple and heterogeneous of Illinois at Urbana-Champaign) media in a principled manner, presentation focused on recognizing

This article describes FAQ FINDER, a natural language question-answering system that uses files of frequently asked questions as its knowledge base. Unlike AI question-answering systems that focus on the generation of new answers, FAQ FINDER retrieves existing ones found in frequently asked question files. Unlike information-retrieval approaches that rely on a purely lexical metric of similarity between query and document, FAQ FINDER uses a semantic knowledge base (WORDNET) to improve its ability to match question and answer. We include results from an evaluation of the system's performance and show that a combination of semantic and statistical techniques works better than any single approach.