A Causal View on Robustness of Neural Networks

  • Cheng Zhang ,
  • Kun Zhang ,
  • Yingzhen Li

2020 Neural Information Processing Systems |

PDF | PDF | Publication | Publication | Publication | Publication | Publication | Publication | Publication | Publication

We present a causal view on the robustness of neural networks against input manipulations, which applies not only to traditional classification tasks but also to general measurement data. Based on this view, we design a deep causal manipulation augmented model (deep CAMA) which explicitly models the manipulations of data as a cause to the observed effect variables. We further develop data augmentation and test-time fine-tuning methods to improve deep CAMA’s robustness. When compared with discriminative deep neural networks, our proposed model shows superior robustness against unseen manipulations. As a by-product, our model achieves disentangled representation which separates the representation of manipulations from those of other latent causes.