Weakly Supervised Invariant Representation Learning Via Disentangling Known and Unknown Nuisance Factors
This work addresses representation learning challenges for machine learning practitioners by combining disentanglement and invariance, though it appears incremental as it builds on existing concepts like contrastive learning and nuisance factor separation.
The paper tackles the complementary goals of learning disentangled and invariant representations by proposing a framework that uses weakly supervised signals to separate predictive, known nuisance, and unknown nuisance information, while incorporating contrastive methods for invariance. The method outperforms state-of-the-art approaches on four benchmarks and demonstrates improved adversarial defense capabilities without adversarial training.
Disentangled and invariant representations are two critical goals of representation learning and many approaches have been proposed to achieve either one of them. However, those two goals are actually complementary to each other so that we propose a framework to accomplish both of them simultaneously. We introduce a weakly supervised signal to learn disentangled representation which consists of three splits containing predictive, known nuisance and unknown nuisance information respectively. Furthermore, we incorporate contrastive method to enforce representation invariance. Experiments shows that the proposed method outperforms state-of-the-art (SOTA) methods on four standard benchmarks and shows that the proposed method can have better adversarial defense ability comparing to other methods without adversarial training.