Bi-directional Reasoning Based on BAM Neural Networks

Neural network is a type of network that carries out information processing through the interaction of neurons. The storage of knowledge and information shows distributed physical connection of mutual-linking network components. Bi-directional associative memories (BAM) neural network is a type of feedback neural network system of bi-directional stability, which exists simple characteristics that can be achieved by large-scale integrate circuit. The paper described general reasoning tactics and their Characteristics, studied the theoretic gist of reasoning using BAM neural networks, and analyzed the running methods of BAM under different reasoning tactics through example. Finally, the problem of network capacity was discussed and the further investigative direction was pointed out.