Decoder-free Robustness Disentanglement without (Additional) Supervision
This addresses adversarial vulnerability in machine learning models by improving feature preservation, though it appears incremental as it builds on existing disentanglement approaches.
The paper tackles the problem of adversarial training causing accuracy reduction by discarding non-robust features, proposing a method to preserve and disentangle both robust and non-robust features without extra supervision, achieving better disentanglement and preserved accuracy.
Adversarial Training (AT) is proposed to alleviate the adversarial vulnerability of machine learning models by extracting only robust features from the input, which, however, inevitably leads to severe accuracy reduction as it discards the non-robust yet useful features. This motivates us to preserve both robust and non-robust features and separate them with disentangled representation learning. Our proposed Adversarial Asymmetric Training (AAT) algorithm can reliably disentangle robust and non-robust representations without additional supervision on robustness. Empirical results show our method does not only successfully preserve accuracy by combining two representations, but also achieve much better disentanglement than previous work.