Satinder Baveja
Director

The Artificial Intelligence (AI) program at the University of Michigan comprises a multidisciplinary group of researchers conducting theoretical, experimental, and applied investigations of intelligent systems. Current projects include research in rational decision making, distributed systems of multiple agents, machine learning, cognitive modeling, design, collaboration technology, natural language processing, real-time and intelligent dynamical control, autonomous and tele-autonomous robotic systems, computer vision, and digital libraries.

Research in the AI laboratory tends to be highly interdisciplinary, building on ideas from computer science, linguistics, psychology, economics, biology, controls, and philosophy. In pursuing this approach, laboratory faculty and students work closely with colleagues throughout the University. This collaborative environment, coupled with our unusually diverse perspectives leads to a valuable interchange of ideas within and across research groups.

The long-term goal of much research at the AI laboratory is to develop autonomous agents capable of behaving effectively in physical and software environments. Effective behavior requires integration of many capabilities. Research toward this goal aims to construct agents that can:

• use and direct their sensory and communicative mechanisms to gather information about their environments and each other;
• represent and reason about the world they inhabit;
• plan actions that are effective despite being undertaken in incompletely understood and dynamically changing environments;
• reason about how best to balance competing goals of choosing actions that are effective and safe, while doing so cheaply and quickly;
• coordinate their actions to cooperatively achieve goals and allocate their resources effectively;
• understand instructions or other communication from humans;
• improve their performance by learning from experience.

Additional research at the laboratory focuses on methods for designing agent interaction environments, predicting agent behavior, and tools for evaluating effective agents.

Often the significance of breakthroughs in theory and in the laboratory are best evaluated in the context of real-world applications. Among the various applications currently explored by members of the AI Laboratory are digital libraries, simulated environments for training people, user interfaces to complex automation systems, mobile robotics for nuclear reactor maintenance, intelligent transportation systems, Internet auctions, information systems for K-12 education and computer games.

The faculty and graduate students of the Artificial Intelligence Laboratory are housed in the Computer Science and Engineering Building, on the University's North Campus. Despite their varied interests, the members shared sense of adventurous research draws them together. This sense of common purpose is facilitated by the Lab's emphasis on moving graduate students into active research as soon as possible.

The AI Lab utilizes a variety of UNIX, Macintosh, and Intel-based workstations to support the research and activities of its faculty, students and staff. High speed fiber-optic and FastEthernet communications networks tie the Lab equipment seamlessly into the department and college computing environments. Three high-speed laser printers and a Tektronix color laser printer make hardcopy printouts look their best. The Lab has light-controlled areas for computer vision and robotics work as well as complete video editing facilities for production of demonstration tapes and presentations. A full-time support staff is available to assist members of the Lab with problems in hardware and software they may encounter. In addition, Lab and Departmental computing support staff operate jointly to make computing support available 24 hours a day.