作者: John A. Campbell , Niladri Chatterjee
DOI:
关键词:
摘要: The non-algorithmic nature of traditional knowledge-based (KB) reasoning approaches makes these techniques open-ended. As a result, KB appear unsuitable for real-time applications as is evident from the small number appearances in domains. technique knowledge interpolation potential remedy shortcoming. Intuitively, imitates numerical-analysis interpolation, derives solutions an unknown problem some already known values, and thereby avoids extensive searches base. extra inherent advantage that it gives computation more predictable algorithmic character. Hence not only can computation's temporal requirements be estimated, but also themselves may reduced significantly. Effective application this needs to answer two questions: what are prerequisites such techniques, possible ways application. This paper studies both issues implementational point view.