Research map / Label-efficient and robust learning / Recognition and long-tail
On Group Sufficiency Under Label Bias
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Characterizes fair classification under group-dependent label noise (label bias) and studies when learning with respect to true labels is feasible from biased observed labels.
Problem
Real-world classification datasets contain label bias: observed labels differ systematically from true labels at group-dependent rates, so models trained on them can show performance disparities across demographic groups.
Approach
Characterizes the problem of learning fair classifiers with respect to the underlying ground-truth labels when only group-biased observed labels are available, using a notion of group sufficiency under label bias.
Closest papers by meaning
- Learning Fair Classifiers With Partially Annotated Group Labels (CVPR 2022 · similarity 0.91)
- Fair Selective Classification Via Sufficiency (ICML 2021 · similarity 0.91)
- Mitigating Label Bias in Machine Learning: Fairness through Confident Learning (AAAI 2024 · similarity 0.90)
- Fairness without Harm: An Influence-Guided Active Sampling Approach (NeurIPS 2024 · similarity 0.87)
- Subgroups Matter for Robust Bias Mitigation (ICML 2025 · similarity 0.87)
- Towards Fairness under Label Bias in Image Segmentation: Impact, Measurement and Mitigation (NeurIPS 2026 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.