Understanding the role of individual units in a deep neural network

• DOI: 10.1073/pnas.1907375117
• Publication Date: 2020-09-01T20:45:27Z
• Authors: David Bau; Jun-Yan Zhu; Hendrik Strobelt; Agata Lapedriza; Bolei Zhou; Antonio Torralba
• Keywords: Contextual image classification

Deep neural networks excel at finding hierarchical representations that solve complex tasks over large data sets. How can we humans understand these learned representations? In this work, present network dissection, an analytic framework to systematically identify the semantics of individual hidden units within image classification and generation networks. First, analyze a convolutional (CNN) trained on scene discover match diverse set object concepts. We find evidence has many classes play crucial roles in classifying classes. Second, use similar method generative adversarial (GAN) model generate scenes. By analyzing changes made when small sets are activated or deactivated, objects be added removed from output scenes while adapting context. Finally, apply our understanding attacks semantic editing.