# BLADE: Bias-Linked Adaptive DEbiasing

arXiv: https://arxiv.org/abs/2510.04174 · HTML: https://mlsanity.com/paper/2510.04174/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 20 of its 26 references that are in the map.
- First posted on arXiv 5 Oct 2025.

## Builds on (most cited first, 20 of 20)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Arbitrary Style Transfer in Real-time with Adaptive Instance Normalization](https://mlsanity.com/paper/1703.06868/) (2017)
- [Distributionally Robust Neural Networks for Group Shifts: On the Importance of Regularization for Worst-Case Generalization](https://mlsanity.com/paper/1911.08731/) (2019)
- [StarGAN: Unified Generative Adversarial Networks for Multi-Domain Image-to-Image Translation](https://mlsanity.com/paper/1711.09020/) (2017)
- [Generalized Cross Entropy Loss for Training Deep Neural Networks with Noisy Labels](https://mlsanity.com/paper/1805.07836/) (2018)
- [Just Train Twice: Improving Group Robustness without Training Group Information](https://mlsanity.com/paper/2107.09044/) (2021)
- [Environment Inference for Invariant Learning](https://mlsanity.com/paper/2010.07249/) (2020)
- [Learning Not to Learn: Training Deep Neural Networks with Biased Data](https://mlsanity.com/paper/1812.10352/) (2018)
- [Learning De-biased Representations with Biased Representations](https://mlsanity.com/paper/1910.02806/) (2019)
- [Learning Debiased Representation via Disentangled Feature Augmentation](https://mlsanity.com/paper/2107.01372/) (2021)
- [BiaSwap: Removing dataset bias with bias-tailored swapping augmentation](https://mlsanity.com/paper/2108.10008/) (2021)
- [Discover and Mitigate Unknown Biases with Debiasing Alternate Networks](https://mlsanity.com/paper/2207.10077/) (2022)
- [SelecMix: Debiased Learning by Contradicting-pair Sampling](https://mlsanity.com/paper/2211.02291/) (2022)
- [Training Debiased Subnetworks with Contrastive Weight Pruning](https://mlsanity.com/paper/2210.05247/) (2022)
- [Fighting Fire with Fire: Contrastive Debiasing without Bias-free Data via Generative Bias-transformation](https://mlsanity.com/paper/2112.01021/) (2021)
- [Enhancing Intrinsic Features for Debiasing via Investigating Class-Discerning Common Attributes in Bias-Contrastive Pair](https://mlsanity.com/paper/2404.19250/) (2024)
- [A Simple Remedy for Dataset Bias via Self-Influence: A Mislabeled Sample Perspective](https://mlsanity.com/paper/2411.00360/) (2024)
- [Hybrid Sample Synthesis-based Debiasing of Classifier in Limited Data Setting](https://mlsanity.com/paper/2312.08288/) (2023)

## Abstract (arXiv)

Neural networks have revolutionized numerous fields, yet they remain vulnerable to a critical flaw: the tendency to learn implicit biases, spurious correlations between certain attributes and target labels in training data. These biases are often more prevalent and easier to learn, causing models to rely on superficial patterns rather than task-relevant features necessary for generalization. Existing methods typically rely on strong assumptions, such as prior knowledge of these biases or access to bias-conflicting samples, i.e., samples that contradict spurious correlations and counterbalance bias-aligned samples, samples that conform to these spurious correlations. However, such assumptions are often impractical in real-world settings. We propose BLADE ({B}ias-{L}inked {A}daptive {DE}biasing), a generative debiasing framework that requires no prior knowledge of bias or bias-conflicting samples. BLADE first trains a generative model to translate images across bias domains while preserving task-relevant features. Then, it adaptively refines each image with its synthetic counterpart based on the image's susceptibility to bias. To encourage robust representations, BLADE aligns an image with its bias-translated synthetic counterpart that shares task-relevant features but differs in bias, while misaligning it with samples sharing the same bias. We evaluate BLADE on multiple benchmark datasets and show that it significantly outperforms state-of-the-art methods. Notably, it exceeds the closest baseline by an absolute margin of around 18% on the corrupted CIFAR-10 dataset under the worst group setting, establishing a new benchmark in bias mitigation and demonstrating its potential for developing more robust deep learning models without explicit supervision.