An Interactive Node-Link Visualization of Convolutional Neural Networks

By Adam W. Harley. 2015

126 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Jun Zhu
2. Chongxuan Li
3. Jiaxin Shi
4. Mengchen Liu
5. Shixia Liu
6. Fred Hohman
7. Xiting Wang
8. Z. Li

In Computer Vision and Pattern Recognition

1. Jun Zhu
2. Xiaoming Liu
3. Chongxuan Li
4. Amin Jourabloo
5. Mengchen Liu
6. Fred Hohman
7. Z. Li
8. Jiaxin Shi

Citations in the corpus

• Towards Better Analysis of Deep Convolutional Neural Networks (2016, background): “More relevant to our work is to visualize neural networks as node-link diagrams [21, 56].”
• CNN Explainer: Learning Convolutional Neural Networks with Interactive Visualization (2020, methodology): “In recent years, many visualization systems have been developed for deep learning, but very few are designed for non-experts [20, 29, 44, 50], as surveyed in [23].”
• ActiVis: Visual Exploration of Industry-Scale Deep Neural Network Models (2017, methodology): “While many existing deep learning visualization tools support instance-based exploration [10, 14, 18, 35, 38], ACTIVIS is the first tool that simultaneously supports instance- and subset-based exploration of the deep neural network models.”
• Understanding convolutional neural networks with a mathematical model (2016, background): “nsor analysis [5], generative modeling [6], relevance propagation [7], Taylor decomposition [8], etc. Another popular topic along this line is on the visualization of lter responses at various layers [9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26]. This research attempts to a”
• Under the Hood of Neural Networks: Characterizing Learned Representations by Functional Neuron Populations and Network Ablations (2020, background): “Aiming to look inside of neural networks, a number of tools have been developed to visualize network activity in response to specific inputs (Harley, 2015) or in response to varying hyper parameters (Smilkov et al.”
• Visual Analytics in Deep Learning: An Interrogative Survey for the Next Frontiers (2018, unknown): “Another example is 3D visualization of a CNN trained on the classic MNIST dataset1 that shows the convolution windows and activations on images that the user draws by hand [37].”
• Visualizations of Deep Neural Networks in Computer Vision: A Survey (2017, methodology): “Harley [26] 2015 User feedback integration Pixel displays, Node-Link-Diagram Recognition CNNs MNIST”
• Skin Cancer Detection: A Review Using Deep Learning Techniques (2021, background): “CNN’s hidden layers consist of convolution layers, nonlinear p oli g layers, and fully connected layers [29].”
• DeepVID: Deep Visual Interpretation and Diagnosis for Image Classifiers via Knowledge Distillation (2019, background): “Conversely, the example-based solutions focus on interpreting the behavior of DNNs with specific data instances as inputs [22], [23].”
• Towards better analysis of machine learning models: A visual analytics perspective (2017, background): “mation of the networks. As a result, they fail to provide a comprehensive understanding of the roles of dierent neurons in dierent layers and the interactions between them. Network-based techniques [23, 24, 25] solve this problem by displaying the network topology. These techniques usually repre5 Fi”
• Do Convolutional Neural Networks Learn Class Hierarchy? (2017, background): “This has been shown useful to in-spect how the network classifies a selected or user-generated sample [26, 67].”