XAI-papers/README.md

# Papers on Explainable Artificial Intelligence

This is an on-going attempt to consolidate all interesting efforts in the area of understanding / interpreting / explaining / visualizing machine learning models.

---------------------------------------

# GUI tools
* Deep Visualization Toolbox. _Yosinski et al. 2015_ [code](https://github.com/yosinski/deep-visualization-toolbox) | [pdf](http://yosinski.com/deepvis)

# Surveys
* Methods for Interpreting and Understanding Deep Neural Networks [pdf](https://arxiv.org/pdf/1706.07979.pdf)
* The Mythos of Model Interpretability [pdf](https://arxiv.org/abs/1606.03490)
* Towards A Rigorous Science of Interpretable Machine Learning [pdf](https://arxiv.org/pdf/1702.08608.pdf)
* Visualizations of Deep Neural Networks in Computer Vision: A Survey (Seifert et al. 2017) [pdf](https://link.springer.com/chapter/10.1007/978-3-319-54024-5_6)
* How convolutional neural network see the world - A survey of convolutional neural network visualization methods (Qin et al. 2018) [pdf](https://arxiv.org/abs/1804.11191)
* A brief survey of visualization methods for deep learning models from the perspective of Explainable AI (Chalkiadakis 2018) [pdf](https://www.macs.hw.ac.uk/~ic14/IoannisChalkiadakis_RRR.pdf)

# Feature Visualization
## Activation Maximization
* Visualizing higher-layer features of a deep network. _Erhan et al. 2009_ [pdf](https://www.researchgate.net/publication/265022827_Visualizing_Higher-Layer_Features_of_a_Deep_Network)
* Synthesizing the preferred inputs for neurons in neural networks via deep generator networks. _Nguyen et al. 2016_ [code](https://github.com/Evolving-AI-Lab/synthesizing) | [pdf](anhnguyen.me/project/synthesizing)
* Plug and Play Generative Networks. _Nguyen et al. 2017_ [pdf](anhnguyen.me/project/ppgn/) | [code](https://github.com/Evolving-AI-Lab/ppgn)
* Feature Visualization. _Olah et al. 2017_ [url](https://distill.pub/2017/feature-visualization)

## Segmentation Masks
* Object Detectors Emerge in Deep Scene CNNs. Zhou et al. 2015 [pdf](https://arxiv.org/abs/1412.6856)
* Network Dissection: Quantifying Interpretability of Deep Visual Representations. Bau et al. 2017 [url](http://netdissect.csail.mit.edu/) | [pdf](http://netdissect.csail.mit.edu/final-network-dissection.pdf)
* Net2Vec: Quantifying and Explaining how Concepts are Encoded by Filters in Deep Neural Networks. Fong & Vedaldi 2018 [pdf](https://arxiv.org/abs/1801.03454)

## Heatmaps
* Learning how to explain neural networks: PatternNet and PatternAttribution [pdf](https://arxiv.org/abs/1705.05598)
* A Theoretical Explanation for Perplexing Behaviors of Backpropagation-based Visualizations [pdf](https://arxiv.org/abs/1805.07039)
* A Taxonomy and Library for Visualizing Learned Features in Convolutional Neural Networks [pdf](https://arxiv.org/pdf/1606.07757.pdf)

### Layer-wise Backpropagation
* Beyond saliency: understanding convolutional neural networks from saliency prediction on layer-wise relevance propagation [pdf](https://arxiv.org/abs/1712.08268)
* Explaining NonLinear Classification Decisions With Deep Tayor Decomposition [pdf](https://arxiv.org/abs/1512.02479)

## Inverting Neural Networks
* Understanding Deep Image Representations by Inverting Them [pdf](https://arxiv.org/abs/1412.0035)
* Inverting Visual Representations with Convolutional Networks [pdf](https://arxiv.org/abs/1506.02753)
* Neural network inversion beyond gradient descent [pdf](http://opt-ml.org/papers/OPT2017_paper_38.pdf)

## Bayesian approaches

* Yang, S. C. H., & Shafto, P. Explainable Artificial Intelligence via Bayesian Teaching. NIPS 2017 [pdf](http://shaftolab.com/assets/papers/yangShafto_NIPS_2017_machine_teaching.pdf)

## Distilling DNNs into more interpretable models
* Interpreting CNNs via Decision Trees [pdf](https://arxiv.org/abs/1802.00121)
* Distilling a Neural Network Into a Soft Decision Tree [pdf](https://arxiv.org/abs/1711.09784)

## DNNs that learn to explain
* Deep Learning for Case-Based Reasoning through Prototypes [pdf](https://arxiv.org/pdf/1710.04806.pdf)
* Unsupervised Learning of Neural Networks to Explain Neural Networks [pdf](https://arxiv.org/abs/1805.07468)

## Others
* Understanding Deep Architectures by Interpretable Visual Summaries [pdf](https://arxiv.org/pdf/1801.09103.pdf)