This paper presents our experience in designing and implementing a data/knowledge base management testbed. The testbed consists of two layers, the knowledge manager and the database management system, with the former at the top. The testbed is based on the logic programming paradigm, wherein data, knowledge, and queries are all expressed as Horn clauses. The knowledge manager compiles pure, function-free Horn clause queries into embedded-SQL programs, which are executed by the database management system to produce the query results. The database management system is a commercial relational database system and provides storage for both rules and facts. First, the testbed architecture and major data structures and algorithms are described. Then, several preliminary tests conducted using the current version of the testbed and the conclusions from the test results are presented. The principal contributions of this work have been to unify various concepts, both previously published and new ones we developed, into a real system and to present several insights into data/knowledge base management system design gleaned from the test results and our design and implementation experience.

This paper analyzes potential optical architectures for AI applications (such as knowledge-based systems). Our goal was to investigate architectures most suitable for implementation completely in optics. While optical computing appears to hold much promise because of its inherent parallelism and speed, constructing a symbolic processor or even a general purpose computer in optics requires examining many issues never before addressed. This paper presents these issues and discusses those architectures which appear most feasible in optics. We take into account fundamental physical limitations as well as the state-of-the-art optical device research. We conclude that, unlike in electronics, large-grained parallelism is not suitable for implementation in optics. We also find that functional languages, rather than logic languages, are better candidates for optics. Finally, we show that implementing an optical symbolic processor warrants the need for a real, or at least an emulated, addressable memory in optics.