Neural network logic system

摘要: A novel neural network implementation for logic systems has been developed. The can determine whether a particular system and knowledge base are self-consistent, which be difficult problem more complex systems. Through hardware using parallel computation, valid solutions may found rapidly than could done with previous, software-based implementations. This is particularly suited use in large, real-time problems, such as expert testing the consistency of programmable process controller, an integrated circuit design, or "expert system." also used "inference engine," i.e., to test validity logical expression context given base, search all solutions, consistent truth values have "clamped" true false. many different types systems: those based on conventional "truth table" logic, maintenance system, other "justifications" corresponding permanently hard-wired by manufacturer, supplied user, either reversibly irreversibly.