Training Method of Causal Extensional Model of Boltzmann Machines
Qingsong Peng · 2023
Boltzmann machine is a kind of probability graphical model, which includes visible layer and hidden layer, and the nodes are fully connected. But this fully connected structure may not be perfect. It is not necessarily to connect or not connect the visible layer as well as the hidden layer. There should be a trade-off between the two, that is, the nodes in the visible layer and the hidden layer are partially connected. According to the restricted Boltzmann machine causality extensional model, the fully connected Boltzmann machine model is introduced. Starting from the training of Boltzmann machine model, this paper introduces two training methods which shows the application solution of the new model.