Efficient debiasing with contrastive weight pruning

Geon Yeong Park, Sang Wan Lee, Jong Chul Ye

Keywords: Debiasing, spurious correlation, pruning

Abstract:

Neural networks are often biased to spuriously correlated features that provide misleading statistical evidence that does not generalize. This raises a fundamental question: "Does an optimal unbiased functional subnetwork exist in a severely biased network? If so, how to extract such subnetwork?" While few studies have revealed the existence of such optimal subnetworks with the guidance of ground-truth unbiased samples, the way to discover the optimal subnetworks with biased training dataset is still unexplored in practice.
To address this, here we first present our theoretical insight that alerts potential limitations of existing algorithms in exploring unbiased subnetworks in the presence of strong spurious correlations. We then further elucidate the importance of bias-conflicting samples on structure learning. Motivated by these observations, we propose a Debiased Contrastive Weight Pruning (DCWP) algorithm, which probes unbiased subnetworks without expensive group annotations. Experimental results demonstrate that our approach significantly outperforms state-of-the-art debiasing methods despite its considerable reduction in the number of parameters.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Deep Learning and representational learning