Parallel Knowledge-Based Information Retrieval on the NON-VON Machine
NON-VON (Shaw, 1979; Shaw et al. 1981) is a highly parallel machine adapted to the extremely efficient support of certain operations that appear central to a wide range of large-scale, knowledge-based information processing tasks. In order to demonstrate the utility of the NON-VON architecture for large-scale Artificial Intelligence applications, we have implemented a simple knowledge-based retrieval system, with the primitive NON-VON machine instructions emulated in software. The system compares KRL-like descriptions (Bobrow and Winograd, 1977) with the contents of what would be a very large database, retrieving all "matching" descriptions the criteria for description-matching require deductive inference over a domain-specific knowledge base. This paper describes the essential mechanisms in our experimental knowledge-based retrieval system.
- cucs-021-81.pdf application/pdf 1.9 MB Download File
More About This Work
- Academic Units
- Computer Science
- Department of Computer Science, Columbia University
- Columbia University Computer Science Technical Reports, CUCS-021-81
- Published Here
- October 21, 2011
This research was supported in part by the Defense Advanced Research Projects Agency under contracts MDA903-77-C-0322 and N00039-80-G-0132.