Research

My previous research contribution can be classified into the four categories:

In production systems, I first proposed parallel rule firing with Sal Stolfo, when researchers mainly worked on parallel rule matching. I viewed production systems as a collection of individual concurrent activities, invented compile/run time algorithms to guarantee serializablility of rule firings, and actually developed parallel firing systems. I then extended parallel firing systems to distributed rule firing and introduced organizational self-design for adapting to environmental changes with Les Gasser. I was a member of the NTT Knowledge Base Management System Project, and developed an intelligent CAL using production systems, which becomes a key product of a NTT group company. The series of papers have been published at various conferences including IEEE CAIA, IEEE TAI, AAAI, IJCAI and appeared as four IEEE TKDE transaction papers. I was invited to give a talk at the first International Conference on Multiagent Systems (ICMAS95).

In multiagent search, since production systems are reactive, I tried to introduce deliberation in multiagent problem solving. I initiated an agent research group in NTT, and start creating computational algorithms for multiagent systems, when researchers mainly focused on conceptual works. I worked on path finding problems and constraint satisfaction problems, the two major search problems in AI. For path finding problems, I extended realtime search to be capable to utilize and improve previous experiments, to adapt to the dynamically changing goals with Richard Korf, and to cooperatively solve problems with other problem solvers. For constraint satisfaction, I worked with Makoto Yokoo and created a new problem called distributed constraint satisfaction, which has been widely accepted in this field. Papers have been published at conferences including IEEE ICDCS, AAAI, IJCAI, and appeared as two IEEE TKED/TPAMI transaction papers. I am a co-author of a multiagent search chapter for the first textbook of multiagent systems, which was published from the MIT press.

In community computing, I created a new application field for autonomous agents and multiagent systems. I realized a paradigm shift in computing metaphors: from team to community. Given that the team metaphor has created research fields like groupware and cooperative agents, the community metaphor will generate new research field. I proposed a concept of communityware to support the process of organizing diverse and amorphous groups of people, while groupware mainly addressed the collaborative work of already-organized people. In other words, compared to groupware studies, he focused on an earlier stage of collaboration: group formation from a wide variety of people. My team developed mobile assistants and tried out them at international conference ICMAS 1996 with 100 PDAs with wireless phones. This work was done with Yoshiyasu Nishibe. I also worked on a 3D interaction space called FreeWalk/Q with Hideyuki Nakanishi, and applied it to Digital City Kyoto. I published three LNCS proceedings and created a network among digital cities in Amsterdam, Helsinki, Seattle, Shanghai and Kyoto.

In intercultural collaboration, I initiated an Intercultural Collaboration Experiments (ICE) with Chinese, Korean, Malaysian colleagues in 2002, a year after 9.11. In 2006, I began the Language Grid project to create a language service infrastructure on the Internet. Basic software for the Language Grid has been studied and developed at the National Institute of Information and Communications Technology (NICT). For trial operation, however, Department of Social Informatics, Graduate School of Informatics, Kyoto University takes on the role as the Language Grid Operator. So far, 140 groups from 18 countries join the Language Grid to share more than 180 language services.