English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Dataset biases are notoriously detrimental to model robustness andgeneralization. The identify-emphasize paradigm appears to be effective indealing with unknown biases. However, we discover that it is still plagued bytwo challenges: A, the quality of the identified bias-conflicting samples isfar from satisfactory; B, the emphasizing strategies only produce suboptimalperformance. In this paper, for challenge A, we propose an effectivebias-conflicting scoring method (ECS) to boost the identification accuracy,along with two practical strategies -- peer-picking and epoch-ensemble. Forchallenge B, we point out that the gradient contribution statistics can be areliable indicator to inspect whether the optimization is dominated bybias-aligned samples. Then, we propose gradient alignment (GA), which employsgradient statistics to balance the contributions of the mined bias-aligned andbias-conflicting samples dynamically throughout the learning process, forcingmodels to leverage intrinsic features to make fair decisions. Furthermore, weincorporate self-supervised (SS) pretext tasks into training, which enablemodels to exploit richer features rather than the simple shortcuts, resultingin more robust models. Experiments are conducted on multiple datasets invarious settings, demonstrating that the proposed solution can mitigate theimpact of unknown biases and achieve state-of-the-art performance.

Delving into Identify-Emphasize Paradigm for Combating Unknown Bias