1. Understanding individual decisions of cnns via contrastive backpropagation;gu;Asian Conference on Computer Vision,2018
2. Attention is all you need;vaswani;Advances in neural information processing systems,2017
3. Imagenet-trained cnns are biased towards texture; increasing shape bias improves accuracy and robustness;geirhos;ArXiv Preprint,2018
4. Instance normalization: The missing ingredient for fast stylization;ulyanov;ArXiv Preprint,2016
5. Visualization of Supervised and Self-Supervised Neural Networks via Attribution Guided Factorization