Understanding Convolutional Networks with APPLE : Automatic Patch Pattern Labeling for Explanation

2018/02/11 by Sandeep Konam, Ian Quah, Konam, Sandeep +5 · 4 citations
Computer Science · Mathematics · #Artificial intelligence #Computer science #Computers and Society (cs.CY) #Convolutional neural network #Explainable Artificial Intelligence (XAI) #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Machine Learning and Data Classification #Natural language processing #Topic Modeling #cs.CY #cs.LG #stat.ML

paper · pdf · doi:10.48550/arxiv.1802.03675

published in arXiv (Cornell University) (Cornell University) · AAAI/ACM Conference on AI, Ethics, and Society

arxiv created 2018/02/11 · openalex publication_date 2018/02/11 · arxiv updated 2018/02/13 · openalex created_date 2018/02/23 · openalex updated_date 2026/07/28

Abstract

With the success of deep learning, recent efforts have been focused on analyzing how learned networks make their classifications. We are interested in analyzing the network output based on the network structure and information flow through the network layers. We contribute an algorithm for 1) analyzing a deep network to find neurons that are 'important' in terms of the network classification outcome, and 2)automatically labeling the patches of the input image that activate these important neurons. We propose several measures of importance for neurons and demonstrate that our technique can be used to gain insight into, and explain how a network decomposes an image to make its final classification.

Citations

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• ImageNet Large Scale Visual Recognition Challenge
• Towards Vision-Based Deep Reinforcement Learning for Robotic Motion Control
• Striving for Simplicity: The All Convolutional Net
• Target-driven Visual Navigation in Indoor Scenes using Deep Reinforcement Learning
• Attentive Explanations: Justifying Decisions and Pointing to the Evidence
• Generating Visual Explanations
• Learning Deep Features for Discriminative Localization
• Context-aware Captions from Context-agnostic Supervision

Related

• A Peek Into the Hidden Layers of a Convolutional Neural Network Through a Factorization Lens
• How You See Me
• Neuron ranking -- an informed way to condense convolutional neural networks architecture
• Shap-CAM: Visual Explanations for Convolutional Neural Networks based on Shapley Value
• Convolutional Neural Networks: