A Logic Connectionist Approach To Self-organized Associative Memory

Much of current research on self-organizod associative memory is based on the sum-of-weights-and-threshold neuron models originally from McCulloch and Pitts. This paper briefly presents a different approach which is based on programmable logic devices similar to the random access memory (RAM). The logic neuron models are based on a logic truth-table with a probabilistic element. The problems of the deterministic logic connectionist systems are described. It is also shown in this paper how the probabilistic logic neural network tackles the problems of the deterministic systems. The current research works on the logic connectionist approach are mentioned.