05-11 (CVPR 2015) Deep neural networks are easily fooled:High confidence predictions for unrecognizable images
05-11 (2017) Paying more attention to attention:Improving the performance of convolutional neural networks via attention transfer