秩引导伪偏差学习用于鲁棒黑盒适配
Rank-Guided Pseudo-Bias Learning for Robust Black-Box Adaptation
提出一种无需访问骨干网络参数和人口统计标注的黑盒去偏框架，通过轻量适配器放大潜在虚假方向、无监督聚类生成伪偏差标签，并结合监督对比对齐与自适应边界净化表示，可直接用于冻结视觉基础模型的下游公平性提升。
原文依据 a fully black-box debiasing framework that requires neither access to backbone parameters nor demographic annotations
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论预训练视觉编码器常被用作冻结的黑盒特征提取器，但它们往往继承虚假相关性，对代表性不足的群体造成不成比例的伤害。我们提出PLD-Debias，一个完全黑盒的去偏框架，既不需要访问骨干网络参数，也不需要人口统计标注。该方法集成了三个组件：（1）秩正则化放大，一种轻量适配器，用于放大潜在的虚假方向；（2）无监督伪偏差归纳，对放大后的特征进行聚类以推断高保真代理偏差标签；（3）偏差引导精炼，结合监督对比对齐与聚类感知的自适应边界来净化表示并均衡决策边界。我们从理论上证明这些组件共同收紧了虚假相关下的最差组风险界。实验上，PLD-Debias在CelebA、Waterbirds和CMNIST上取得了最优的最差组准确率，相比先前黑盒方法提升了3-5个点，同时保持平均准确率。值得注意的是，我们的伪偏差标签与真实偏差标注的保真度超过90%，从而在无需人口统计监督的情况下实现了oracle级别的鲁棒性。结果表明，通过一个即插即用的分类器适配器，可以为任何冻结的基础模型同时实现公平性与效用。
英文原文摘要 1387 字符
Pretrained vision encoders are widely used as frozen, black-box feature extractors, yet they often inherit spurious correlations that disproportionately harm underrepresented groups. We introduce PLD-Debias, a fully black-box debiasing framework that requires neither access to backbone parameters nor demographic annotations. Our method integrates three components: (1) Rank-Regularized Amplification, a lightweight adapter that exaggerates latent spurious directions; (2) Unsupervised Pseudo-Bias Induction, which clusters amplified features to infer high-fidelity proxy bias labels; and (3) Bias-Guided Refinement, combining supervised contrastive alignment with cluster-aware adaptive margins to purify representations and equalize decision boundaries. We theoretically show that these components jointly tighten a worst-group risk bound under spurious correlations. Empirically, PLD-Debias achieves state-of-the-art worst-group accuracy across CelebA, Waterbirds, and CMNIST, improving performance by 3-5 points over prior black-box methods while maintaining average accuracy. Remarkably, our pseudo-bias labels align with ground-truth bias annotations at over 90% fidelity, enabling oracle-level robustness without demographic supervision. Our results demonstrate that fairness and utility can be achieved through a plug-and-play classifier adapter for any frozen foundation model.