Dataset biases are notoriously detrimental to model robustness and generalization. The identify-emphasize paradigm appears to be effective in dealing with unknown biases. However, we discover that it is still plagued by two challenges: A, the quality of the identified bias-conflicting samples is far from satisfactory; B, the emphasizing strategies only produce suboptimal performance. In this paper, for challenge A, we propose an effective bias-conflicting scoring method (ECS) to boost the identification accuracy, along with two practical strategies -- peer-picking and epoch-ensemble. For challenge B, we point out that the gradient contribution statistics can be a reliable indicator to inspect whether the optimization is dominated by bias-aligned samples. Then, we propose gradient alignment (GA), which employs gradient statistics to balance the contributions of the mined bias-aligned and bias-conflicting samples dynamically throughout the learning process, forcing models to leverage intrinsic features to make fair decisions. Furthermore, we incorporate self-supervised (SS) pretext tasks into training, which enable models to exploit richer features rather than the simple shortcuts, resulting in more robust models. Experiments are conducted on multiple datasets in various settings, demonstrating that the proposed solution can mitigate the impact of unknown biases and achieve state-of-the-art performance.

Models notoriously suffer from dataset biases which are detrimental to robustness and generalization. The identify-emphasize paradigm shows a promising effect in dealing with unknown biases. However, we find that it is still plagued by two challenges: A, the quality of the identified bias-conflicting samples is far from satisfactory; B, the emphasizing strategies just yield suboptimal performance. In this work, for challenge A, we propose an effective bias-conflicting scoring method to boost the identification accuracy with two practical strategies -- peer-picking and epoch-ensemble. For challenge B, we point out that the gradient contribution statistics can be a reliable indicator to inspect whether the optimization is dominated by bias-aligned samples. Then, we propose gradient alignment, which employs gradient statistics to balance the contributions of the mined bias-aligned and bias-conflicting samples dynamically throughout the learning process, forcing models to leverage intrinsic features to make fair decisions. Experiments are conducted on multiple datasets in various settings, demonstrating that the proposed solution can alleviate the impact of unknown biases and achieve state-of-the-art performance.

Deep image classifiers have been found to learn biases from datasets. To mitigate the biases, most previous methods require labels of protected attributes (e.g., age, skin tone) as full-supervision, which has two limitations: 1) it is infeasible when the labels are unavailable; 2) they are incapable of mitigating unknown biases -- biases that humans do not preconceive. To resolve those problems, we propose Debiasing Alternate Networks (DebiAN), which comprises two networks -- a Discoverer and a Classifier. By training in an alternate manner, the discoverer tries to find multiple unknown biases of the classifier without any annotations of biases, and the classifier aims at unlearning the biases identified by the discoverer. While previous works evaluate debiasing results in terms of a single bias, we create Multi-Color MNIST dataset to better benchmark mitigation of multiple biases in a multi-bias setting, which not only reveals the problems in previous methods but also demonstrates the advantage of DebiAN in identifying and mitigating multiple biases simultaneously. We further conduct extensive experiments on real-world datasets, showing that the discoverer in DebiAN can identify unknown biases that may be hard to be found by humans. Regarding debiasing, DebiAN achieves strong bias mitigation performance.

Artificial intelligence systems derive their power in learning to perform their tasks directly from data. As a result, AI systems are at the mercy of their training data and in most cases are strictly forbidden to learn anything beyond what is contained in their training data. Data by itself has some principal problems: It is noisy, nearly never complete, and it is dynamic as it continually changes over time. This noise can manifest in many ways in the data -- it can arise from incorrect labels, incomplete labels or misleading correlations. As a result of these problems with data, most AI systems must be very carefully taught how to make decisions, act or respond in the real world. This'careful teaching' involves three stages.

Artificial intelligence systems derive their power in learning to perform their tasks directly from data. As a result, AI systems are at the mercy of their training data and in most cases are strictly forbidden to learn anything beyond what is contained in their training data. Data by itself has some principal problems: It is noisy, nearly never complete, and it is dynamic as it continually changes over time. This noise can manifest in many ways in the data -- it can arise from incorrect labels, incomplete labels or misleading correlations. As a result of these problems with data, most AI systems must be very carefully taught how to make decisions, act or respond in the real world. This'careful teaching' involves three stages.