Acquisition of state transitions in neural network

Abstract

In the artificial intelligence, the breadth-first search is optimal with uniform cost. But it takes long time to obtain the solution. Neural networks process states transitions in parallel with learning ability. We developed a search procedure of states transition doing the the breadth-first, in the neural network. First, the input pattern states are self-organized in the neural network, which consists of the Kohonen layer followed by the state planning layer. The state planning layer makes lateral connections between cells of transitions. Then, the initial and the target states are given as a problem. The network shows an optimal state transition pathway in the neuron firings. Next, the state transition procedure is developed for the formation of the concept of action planning. Here, as the action planning, an integration between the symbols and the action pattern is carried out in the extended neural network.