Characterising Bias in Compressed Models Popular compression techniques turned out to amplify bias in deep neural networks. ArXiV: https://arxiv.org/abs/2010.03058 #NN #DL #bias 14.5K views08:28
