从偏好平均视角揭示RLHF中程序公平性缺陷，为对齐公平研究提供新警示。
arXiv:2608.10126v1 Announce Type: new Abstract: Reinforcement Learning from Human Feedback (RLHF) aggregates heterogeneous preferences into a single r…
表面公平未必真公平，揭秘大模型推荐器在隐藏输出层的公平性缺口，算法透明度的必修课。
arXiv:2608.08284v1 Announce Type: new Abstract: Fairness audits for LLM-based recommenders have largely focused on observable outputs, implicitly assu…
用双层优化框架同时提升预测精度与公平性，为公平机器学习提供新视角。
arXiv:2607.05098v1 Announce Type: new Abstract: When sensitive attributes are continuous and high-dimensional $-$ demographic score vectors, posterior…
机器人按文化偏好决定救助顺序时，隐藏的道德偏见值得每个人警惕。
arXiv:2606.28345v1 Announce Type: cross Abstract: LLM-governed social robots increasingly decide who receives real-world assistance first. As prioriti…
别让表面效度骗了你——FAccT新方法教你证伪预测算法的判别效度，直击AI验证中的隐蔽陷阱
arXiv:2601.17146v2 Announce Type: replace-cross Abstract: Empirical investigations into unintended model behavior often show that the algorithm is pre…
公平性感知的属性表示如何避免分箱偏差？这篇论文给出了系统性的无偏分箱方法，值得关注。
arXiv:2509.21785v2 Announce Type: replace-cross Abstract: Discretizing raw features into bucketized attribute representations is a popular step before…
因果图未知时如何保证公平？这项研究在变量簇上实现了干预公平，为AI可靠性提供新解法。
arXiv:2602.23611v2 Announce Type: replace-cross Abstract: Algorithmic decisions about individuals require predictions that are not only accurate but a…
图神经网络中几何公平性的新视角：从空间结构出发，避免“相似节点”导致的偏见放大
arXiv:2606.17684v1 Announce Type: cross Abstract: Graph-based learning methods have become increasingly prominent due to their strong performance acro…
揭示大模型推荐中隐藏的品牌偏见与认知操纵，挑战AI推荐系统的公平性。
arXiv:2606.17443v1 Announce Type: new Abstract: Large language models (LLMs) are becoming a major way for consumers to find products, but we do not ye…
提出α-公平保险定价框架，实现公平与效率的可调节连续体，为保险精算中的公平性度量提供新视角。
arXiv:2606.14898v1 Announce Type: new Abstract: Fairness in insurance pricing remains a long-standing and deeply debated puzzle. On one hand, insurers…
论文提出“文化漏斗”概念，揭示数据中缺失的文化元素会导致AI对齐失败，挑战传统对齐思路。
arXiv:2606.13808v1 Announce Type: new Abstract: Current cultural alignment approaches focus on inference-time interventions, assuming models already c…
顶尖会议论文揭示人脸识别系统如何系统性扼杀认知多样性，提出「计算认识论灭绝」的警示概念
arXiv:2606.07628v1 Announce Type: cross Abstract: While the eugenic roots of computer vision are well-documented in critical technology studies, less …
探讨数据不可靠时如何保证公平性审计的鲁棒性，挑战传统方法假设。
arXiv:2506.23033v2 Announce Type: replace Abstract: Fairness audits are a key component of responsible machine-learning deployment. Yet, the reliabili…
斯坦福研究揭示，AI招聘工具按EEOC“五分之四规则”检测出显著种族歧视，系统性拒绝非白人候选人。
Article URL: https://hai.stanford.edu/news/ai-hiring-tools-can-yield-racial-bias-and-systemic-rejection Comments URL: https://news.ycombinator.com/ite…
联邦学习中的公平共形预测新方法FedCF，解决异质数据下的不确定性量化与公平性挑战。
arXiv:2509.22907v2 Announce Type: replace Abstract: Conformal Prediction (CP) is a widely used technique for quantifying uncertainty in machine learni…
如何在零和博弈的反馈机制中确保公平？这篇论文提出了新视角与约束方案。
arXiv:2606.01159v1 Announce Type: new Abstract: We study two-player zero-sum games (TPZSGs) with bandit feedback under fairness constraints requiring …
提出风险厌恶视角下的公平多类分类新框架，兼顾统计公平与极端情况鲁棒性。
arXiv:2509.05771v2 Announce Type: replace-cross Abstract: We develop a new classification framework based on the theory of coherent risk measures and …
用物理中的能量护盾概念解决AI公平性难题，理论创新与算法验证兼备
arXiv:2605.24926v1 Announce Type: new Abstract: Runtime fairness is not a one-time constraint but a dynamic property evaluated over a sequence of deci…
揭示AI评估系统设计中的固有偏见问题
Article URL: https://alokit.substack.com/p/your-ai-evaluation-is-biased-by-design Comments URL: https://news.ycombinator.com/item?id=48269500 Points: …
面向选择性标签场景，提出长期公平性框架，解决机器学习中的动态公平问题。
arXiv:2605.22291v1 Announce Type: new Abstract: Long-term fairness algorithms aim to satisfy fairness beyond static and short-term notions by accounti…