[ Home | Lab | New Students | Courses | Research | Publications | Activities ]
[ Research Interests | Research Sponsorship | Published Research ]

Research Interests

The best way to learn about my research is from my papers and from the dissertations and theses I have supervised.

However, you can get a quick and dirty feel for my take on the world from my columns, several of which I authored from 1997 to 2002. It is surprising to me that many of the old topics (and my takes on them) have persisted over two decades.

My lifelong program of research can be summed up as discovering, modeling, studying, and applying high-level abstractions for computing.

Computing is in the midst of a paradigm shift from individual computations to interaction. Current techniques are woefully inadequate for network-based applications such as ubiquitous information access and e-commerce, which involve a number of heterogeneous (or independently designed) and autonomous (or independently operated) subsystems. The metaphor of interaction offers a natural, powerful conceptual basis for designing solutions for these applications. I take my inspiration from current developments in databases, artificial intelligence, and distributed computing.

Agents are a popular metaphor for autonomous, persistent computations that perceive, reason, act, and communicate. Agents interact to form multiagent systems, which promise a natural means for building complex computing systems in large, dynamically changing, networked environments.

The motivation for my research program is practical, but there are major implications on problems and approaches studied within computer science. I was among the first researchers to study theories of agents and communication, apply agent techniques in workflow management, and use ontologies for heterogeneous database access. My current research falls into two major themes, emphasizing theory and applications, respectively.

My work is inherently interdisciplinary. It synthesizes insights from different subareas of computer science, primarily distributed artificial intelligence and distributed computing. It also involves disciplines such as linguistics, philosophy, organizational theory, and law. Being interdisciplinary increases the startup cost for research, but helps make the results broadly applicable and thus more valuable.

I have made a conscious effort to reach a wide audience through my research. Besides publishing in technical specialty venues, I have sought to publish in wider-circulating forums that address computer scientists in general.