Device, method and program for erasing select contributing layers of a neural network
Abstract:
A learning device (10) calculates, for each layer in a multilayer neural network, a degree of contribution indicating a degree of contribution to an estimation result of the multilayer neural network, and selects a to-be-erased layer on the basis of the degree of contribution of each layer. The learning device (10) erases the to-be-erased layer from the multilayer neural network, and learns the multilayer neural network from which the to-be-erased layer has been erased.
Information query
Patent Agency Ranking
0/0