Schedule of Papers and Events

In the 1970s, AI leaders believed we would soon see all kinds of intelligent agents, agents that would easily beat grand masters at chess, speak natural language fluently and continually learn new things. These problems turned out to be much harder than they thought. Subfields formed within AI to pursue specialized research in robotics, machine vision, natural language and so on. Speech research was joined with natural language to work towards spoken language. Happily, in past 25 years, these subfields have made significant progress, to the point that there are now commercially available robots of all shapes and sizes, speech recognition engines, and vision systems taht can actually find and sometimes correctly identify faces, gaze, arms, and simple objects.

AI researchers are now putting all these pieces together, either as human-robot interaction (HRI) or, without the mechanical hardware, as on-screen agents (so called embodied conversational agents or ECAs). Building the first generation of such agents has required solving many engineering challenges, and many problems remain. In this talk, I'll give some examples of intriguing embodied intelligent agents, both the physical and on-screen types. However, to be a real success, not only must the engineering get simpler, but there must be some reason for people to want to interact with them. In this talk, I will also look at what some of these reasons might be, drawing on my own work as well as others in HRI and ECAs. 

Candace Sidner is a division scientist at BAE Systems Advanced Information Technologies. She was previously a senior research scientist at Mitsubishi Electric Research Laboratories. She earned her Ph.D. at the MIT Artificial Intelligence Laboratory. Sidner is a fellow and past councilor of AAAI, as well as having served as president of the Association for Computational Linguistics, chair of the 2001 and program cochair of the 2006 International Conference on Intelligent User Interfaces, and cochair of the 2004 SIGdial Workshop on Discourse and Dialogue, and general chair of the 2007 NAACL Human Language Technology conference.

The creative process, for centuries, has been infused with an element of magic, an unexplainability that is supposed to elevate some human minds above the rest - and certainly above mere machines. But creative ideas are, first and foremost, ideas. A fundamental axiom of AI is that the production of ideas can be explained as symbol manipulation and can be reproduced with a symbol manipulation machine. Thus we should be prepared to offer a model of the computational mechanisms by which creative ideas are produced. Some work in psychology and some AI programs suggest mechanisms we will explore as candidates for a model of creativity.

Bruce Buchanan is University Professor of Computer Science Emeritus, and former Professor of Philosophy, Medicine, and Intelligent Systems at the University of Pittsburgh. He is a member of the National Academy of Science Institute of Medicine and has also served as President and Secretary-Treasurer of the AAAI. His main research interests are in machine learning, knowledge-based systems, medical expert systems, and computational biology, and is generally interested in applications of machine learning and artificial intelligence to problems in biology or medicine. His recent research includes: AI approaches to machine learning and discovery, applications of symbolic learning to problems in biology and medicine, and case-based reasoning with application to prediction of protein secondary structure.

Personalized interaction with computer systems can be at odds with privacy since it necessitates the collection of considerable amounts of personal data. Numerous consumer surveys revealed that computer users are very concerned about their privacy online. The collection of personal data is also subject to legal regulations in many countries and states. This talk presents work in the area of Privacy- Enhanced Personalization that aims at reconciling personalization with privacy through suitable human-computer interaction strategies and privacy-enhancing technologies.

Alfred Kobsa is a Professor in the Donald Bren School of Information and Computer Sciences of the University of California, Irvine. Before he was a Director of the Institute for Applied Information Technology (FIT) at the German National Research Center for Information Technology (GMD), and a Professor of Computer Science at the University of Essen, Germany. Dr Kobsa's research lies in the areas of user modeling and personalized systems (with applications in the areas of information environments, expert finders, and user interfaces for disabled and elderly people), privacy, and in information visualization. He is the editor of User Modeling and User-Adapted Interaction, editorial board member of World-Wide Web, Universal Access in the Information Society and Information Technology and Decision Making, and was the founding president of User Modeling Inc. Dr. Kobsa edited several books and authored numerous publications in the areas of user-adaptive systems, human-computer interaction and knowledge representation. He also co-founded a national workshop series and an international conference series in these areas.