Network Pathway Extraction Focusing on Object Level

In this paper, I propose an efficient method of identifying important neurons that are related to object’s concepts by mainly considering the relationship between these and their object concept or class. first quantify activation values among neurons, based on which histograms each neuron generated. Then, obtained clustered identify neurons’ importance. A network-wide holistic approach is also introduced efficiently influential connections reveal pathway a given The as well carefully evaluated sub-network concepts. experimental results MNIST Fashion datasets show effectiveness proposed method.