A Novel Modular Neural Architecture for Rule-based and Similarity-based ReasoningRafal Bogacz, Christophe Giraud-Carrier, A Novel Modular Neural Architecture for Rule-based and Similarity-based Reasoning. Chapter in Hybrid Neural Systems, LNAI 1778. S. Wermter, R. Sun, (eds.), pp. 63–77. March 2000. PDF, 229 Kbytes.
Hybrid connectionist symbolic systems have been the subject of much recent research in AI. By focusing on the implementation of high-level human cognitive processes (e.g., rule-based inference) on low-level, brain-like structures (e.g., neural networks), hybrid systems inherit both the efficiency of connectionism and the comprehensibility of symbolism. This paper presents the Basic Reasoning Applicator Implemented as a Neural Network (BRAINN). Inspired by the columnar organisation of the human neocortex, BRAINN's architecture consists of a large hexagonal network of Hopfield nets, which encodes and processes knowledge from both rules and relations. BRAINN supports both rule-based reasoning and similarity-based reasoning. Empirical results demonstrate promise.