Opening the black box - data driven visualization of neural networks

By Fan-Yin Tzeng, Kwan-Liu Ma. 2005

164 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

In Computer Vision and Pattern Recognition

Citations in the corpus

Explanations of model predictions with live and breakDown packages (2018, background): “Find some examples in Štrumbelj and Kononenko (2011); Tzeng and Ma (2005); Zeiler and Fergus (2014).”
Towards Better Analysis of Deep Convolutional Neural Networks (2016, methodology): “Previous research has also shown that analyzing such connections can help experts understand how a large number of non-linear parts interact with each other [56].”
Peeking Inside the Black Box: Visualizing Statistical Learning With Plots of Individual Conditional Expectation (2013, background): “Tzeng (2005) develops visualization of the layers of neural networks to understand dependencies between the inputs and model outputs and yields insight into classiﬁcation uncertainty.”
Visual Analysis of Hidden State Dynamics in Recurrent Neural Networks (2016, background): “In [24], the authors present a visualization system for feed-forward neural networks with the goal of interpretation, and in [13], the authors give a user-interface for tuning the learning itself.”
Explaining Decision-Making Algorithms through UI: Strategies to Help Non-Expert Stakeholders (2019, methodology): “Examples of the white-box technique include showing probabilities of the nodes for Bayesian networks [4], projection techniques [9] and Nomograms [28] to see the “cut” in the data points for Support Vector Machines, and the visualization of the graph of a neural network [50].”
ActiVis: Visual Exploration of Industry-Scale Deep Neural Network Models (2017, methodology): “Tzeng and Ma [36] was one of the first visualization tools designed for neural network models.”
Visualizing Dataflow Graphs of Deep Learning Models in TensorFlow (2018, background): “Visualizations can improve the transparency and interpretability of the models and help open these “black boxes” [34, 54].”
An Interactive Node-Link Visualization of Convolutional Neural Networks (2015, background): “, [21,22]) typically centers around designing new networks, and making minute adjustments to trained networks, to see how these changes affect performance.”
Towards better analysis of machine learning models: A visual analytics perspective (2017, background): “4 shows the visualization generated by a pioneer network-based technique [17].”
Using sensitivity analysis and visualization techniques to open black box data mining models (2013, methodology): “Regarding the latter approach, some graphical methods were proposed, such as: Hinton and Bond diagrams for NN [9]; showing NN weights and classification uncertainty [31]; and improving the interpretability of kernel-based classification methods [5].”
RuleMatrix: Visualizing and Understanding Classifiers with Rules (2018, methodology): “In pioneering work by Tzeng and Ma [42], a node-linked visualization is used to understand and analyze a trained neural network’s behavior in classifying volume and text data.”
Explanation and Justification in Machine Learning : A Survey Or (2017, methodology): “More recently, visualization techniques have focused on visualizing the hidden states of neural models [Tzeng and Ma, 2005], most notably of Convolutional Neural Nets (CNNs) in image classification [Simonyan et al.”
Visual Interaction with Deep Learning Models through Collaborative Semantic Inference (2019, methodology): “Activation Atlas [12], Classiﬁcation Visualization [15], DeepEyes [65], Feature classiﬁers [8], Tensorboard [84], Weight Visualization [76] Deconvolution [90], LIME [66], Neuron Analysis [20], Rationals [54], Saliency [55], Structured Interpretation [22], Prediction Difference [92] Interactive…”
A Framework for Considering Comprehensibility in Modeling (2016, unknown): “For example, [31] provides visualizations of neural network architectures (internals) while [7] derives visualizable representations of the relationships (sensitivities) between model inputs and outputs.”
Visual Analytics in Deep Learning: An Interrogative Survey for the Next Frontiers (2018, unknown): “be proposed [13], and the trend has continued on in literature.”
LSTMVis: A Tool for Visual Analysis of Hidden State Dynamics in Recurrent Neural Networks (2016, methodology): “[30] present a visualization system for feedforward neural networks with the goal of interpretation, and Kapoor et al.”
Do Convolutional Neural Networks Learn Class Hierarchy? (2017, background): “This has been shown useful to inspect how the network classifies a selected or user-generated sample [26, 67].”