The Design of a Marker Passing Architecture for Knowledge Processing

Knowledge processing is very demanding on computer architectures. Knowledge processing generates subcomputation paths at an exponential rate. It is memory intensive and has high communication requirements. Marker passing architectures are good candidates to solve knowledge processing problems. In this paper, we justify the design decisions made for the Semantic Network Array Processor (SNAP). Important aspects of SNAP are: the instruction set, markers, relations, propagation rules, interconnection network, and granularity. These features are compared to those in NETL and the Connection Machine. 1 Basic Operations in Knowledge Processing The computations that are typical of knowledge processing require the generation of numerous computation paths that all could potentially be followed in parallel. The process of spawning a number of rela tively independent subcomputations, each of which may spawn other subcomputations, is called bifurcation. Bifurcation processes appear to be important...

• Type: Conference
• Year: 1990
• Where: AAAI
• Authors: Wing Lee, Dan I. Moldovan
• Keywords: AAAI 1990, Computation Paths, Intelligent Agents, Knowledge Processing, Knowledge Processing Problems