A Taxonomy and Library for Visualizing Learned Features in Convolutional Neural Networks

Abstract

Over the last decade, Convolutional Neural Networks (CNN) saw a tremendous surge in performance. However, understanding what a network has learned still proves to be a challenging task. To remedy this unsatisfactory situation, a number of groups have recently proposed different methods to visualize the learned models. In this work we suggest a general taxonomy to classify and compare these methods, subdividing the literature into three main categories and providing researchers with a terminology to base their works on. Furthermore, we introduce the FeatureVis library for MatConvNet: an extendable, easy to use open source library for visualizing CNNs. It contains implementations from each of the three main classes of visualization methods and serves as a useful tool for an enhanced understanding of the features learned by intermediate layers, as well as for the analysis of why a network might fail for certain examples.

References in corpus (6)

• Improving neural networks by preventing co-adaptation of feature detectors
• Striving for Simplicity: The All Convolutional Net
• Understanding Neural Networks Through Deep Visualization
• Object Detectors Emerge in Deep Scene CNNs
• Learning Deep Features for Discriminative Localization
• Understanding Deep Image Representations by Inverting Them

Cited by in corpus (14)

• Do Convolutional Neural Networks Learn Class Hierarchy?
• The State of the Art in Enhancing Trust in Machine Learning Models with the Use of Visualizations
• Visual Explanation by Interpretation: Improving Visual Feedback Capabilities of Deep Neural Networks
• Geo-Context Aware Study of Vision-Based Autonomous Driving Models and Spatial Video Data
• Explaining Convolutional Neural Networks using Softmax Gradient Layer-wise Relevance Propagation
• Finding and Visualizing Weaknesses of Deep Reinforcement Learning Agents
• Investigating Saturation Effects in Integrated Gradients
• Stochastic Attraction-Repulsion Embedding for Large Scale Image Localization
• Including Physics in Deep Learning -- An example from 4D seismic pressure saturation inversion
• Deep Epitome for Unravelling Generalized Hamming Network: A Fuzzy Logic Interpretation of Deep Learning
• Towards Human-Understandable Visual Explanations:Imperceptible High-frequency Cues Can Better Be Removed
• Towards the Characterization of Representations Learned via Capsule-based Network Architectures
• Understanding Regularization to Visualize Convolutional Neural Networks
• An Analysis of Human-centered Geolocation