The philosophy behind the research

Information science has developed into knowledge science and its importance will increase more and more in the next century. In the real world, tiny computers of the higher performance have been used by ordinary people in their homes where they also enjoy access to vast amount of information sources scattered all over the world thanks to the internet. In academic society, on the other hand, the research on artificial intelligence is changing from building stand-alone machines which try to solve problems by itself to building intelligent partners which augment human capability of problem solving. This division has been run under the philosophy that it contributes not only to the promotion of knowledge science but also to prosperity of the real world by the feedback of the research results to it in the information era. The major topic here is to investigate Ontological Engineering(Article1 Article2) to establish basic theories and technologies for the next-generation knowledge science. The current research projects include: methodology for ontology design and its representation language, ontology-based next-generation knowledge systems, knowledge sharing and reuse, and intelligent educational systems.

The research map depicts an overview of Mizoguchi lab's research topics on the basis of ontology theory and technology. The basic theories we are developing include semantics of semantic links, functions of an ontology as a meta-model, theories of identity and Part/whole, and the basic technology includes task ontology representation language, methodology of task/domain ontology design as well as its development environment. Ontologies of time, causality, function, behavior, fault, and design tasks are designed as a set of essential building blocks for knowledge systems believing that the next-generation knowledge systems should be built based on well-designed ontology. Intelligent design support systems with functional understanding capability, advanced diagnostic systems and intelligent tutoring/training systems(ITSs) are developed. An ontology-based authoring tool for ITSs is also designed with a tight collaboration with industry. A sophisticated negotiation theory for opportunistic group formation in CSCL(Computer Supported Collaborative Learning) is developed as well as a negotiation ontology which contributes to modeling of collaboration.

In AI research history, we can identify two types of research. One is "Form-oriented research" and the other is "Content-oriented research". The former deals with logic, knowledge representation, etc. and the latter content of knowledge which is critical to modelling in general. Apparently, the former has dominated AI research to date, which does not allow us to say we have successfully conducted AI research and contributed a lot to the real world. Recently, however, "Content-oriented research" has become to gather more attention because a lot of real-world problems to solve such as knowledge reuse, facilitation of agent communication, media integration through understanding, large-scale knowledge bases, etc. require not only advanced formal theories or reasoning methods but also sophisticated treatment of the content of knowledge.

Unfortunately, however, the content-oriented research is not well-established enough to respond to the expectations or to become a branch of a computer science due to the following shortcomings:

1. It tends to be ad-hoc,
2. It does not have a methodology which enables the research results to accumulate, and
3. It lacks firm basic theories and fundamental technologies.

Ontological engineering is what we need to resolve all of these. It eventually provides a basis of building models of all things which computer science is interested in. Ontological engineering gives us sophisticated theories and technologies for explication and utilization of design rationale of a knowledge base, kernel conceptualization of the world of interest, strict definitions of agreed meanings of basic concepts, etc. to enable "accumulation of knowledge" which is indispensable for efficiently modelling the real world.