Representing Knowledge Intelligently: Production Rules, Frames, and Transitional Networks.

摘要

Designs for a production rule interpreter, a frame-based knowledge system, and an augmented transitional network are described. Knowledge represented in the form of production rules (consisting of domain facts and heuristics) is a good way to model the strong data-driven nature of intelligent action. Using this knowledge, production systems make inferences on the system's current understanding of the 'state of the world'. The production rule interpreter described uses a bottom-up approach employing a forward-chaining control strategy. Frames are complex data structures for representing stereotyped objects, events, or situations. For intelligent computer programs requiring this type of information, frame-based knowledge systems minimize redundant information and may be utilized for acquiring new information which is interpreted in terms of concepts acquired through previous experience. The frame-based system described utilizes procedural as well as declarative information. An augmented transitional network is used for natural language as well as declarative information. An augmented transitional network is used for natural language understanding. Keywords: Artificial intelligence; Natural language; Frame-based knowledge; Production system; Augmented networks; Expert systems. (jes)