Towards Better Analysis of Deep Convolutional Neural Networks

Abstract

Deep convolutional neural networks (CNNs) have achieved breakthrough performance in many pattern recognition tasks such as image classification. However, the development of high-quality deep models typically relies on a substantial amount of trial-and-error, as there is still no clear understanding of when and why a deep model works. In this paper, we present a visual analytics approach for better understanding, diagnosing, and refining deep CNNs. We formulate a deep CNN as a directed acyclic graph. Based on this formulation, a hybrid visualization is developed to disclose the multiple facets of each neuron and the interactions between them. In particular, we introduce a hierarchical rectangle packing algorithm and a matrix reordering algorithm to show the derived features of a neuron cluster. We also propose a biclustering-based edge bundling method to reduce visual clutter caused by a large number of connections between neurons. We evaluated our method on a set of CNNs and the results are generally favorable.

arXiv commentSubmitted to VIS 2016

References in corpus (9)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems
• Fast algorithm for detecting community structure in networks
• Caffe: Convolutional Architecture for Fast Feature Embedding
• Understanding Neural Networks Through Deep Visualization
• NodeTrix: Hybrid Representation for Analyzing Social Networks
• Multifaceted Feature Visualization: Uncovering the Different Types of Features Learned By Each Neuron in Deep Neural Networks
• Understanding Intra-Class Knowledge Inside CNN
• Optimal Rectangle Packing: An Absolute Placement Approach

Cited by in corpus (5)

• Towards Interpretable Deep Neural Networks by Leveraging Adversarial Examples
• Improving Visualization Interpretation Using Counterfactuals
• Improving Interpretability of Deep Neural Networks with Semantic Information
• Every Filter Extracts A Specific Texture In Convolutional Neural Networks
• Understanding Convolutional Neural Networks with A Mathematical Model