发表机构
AI Lab, Vrije Universiteit Brussel（人工智能实验室，布鲁塞尔自由大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究基于机器学习的招聘中预训练嵌入的性别偏见问题，通过在合成数据集上评估模型、采用多任务对抗学习框架及多目标模型选择，发现清理性别可减少但未消除泄露，对抗学习能改善公平性，是补充策略。
AI 中文摘要
基于人工智能的招聘系统依赖于在历史简历数据上训练的机器学习模型，存在延续和放大社会偏见的风险。在非结构化简历文本中出现一个关键挑战，即即使去除明确指标，预训练语言模型嵌入仍可能推断出性别等敏感属性。本文在合成FairCVdb数据集上评估了九个预训练嵌入模型，分析其嵌入对申请人评分的信息性以及对性别泄露的敏感性。还使用带有梯度反转的多任务对抗学习框架预测申请人适用性，同时抑制学习表示中的性别信息。最后用基于多目标帕累托前沿的模型选择来平衡预测效用和公平性。实验结果表明，明确的性别清理大幅减少但未消除性别泄露，对抗学习主要改善原始传记的公平性，是一种补充策略而非文本级去偏的替代方法。
英文摘要
AI-based recruitment systems that rely on machine learning models trained on historical CV data, risk perpetuating and amplifying social biases. A key challenge arises in unstructured CV text, where pre-trained language model embeddings may infer sensitive attributes such as gender even after explicit indicators are removed. In this paper, we evaluate nine pre-trained embedding models on the synthetic FairCVdb dataset, analyzing the informativeness of their embeddings for applicant scoring and their susceptibility to gender leakage, on both original and gender-scrubbed biographies. We further use a multi-task adversarial learning framework with gradient reversal to predict applicant suitability while suppressing gender information from learned representations. Finally, we use a multi-objective Pareto-front-based model selection to balance predictive utility and fairness. Our experimental results show that explicit gender scrubbing substantially reduces but does not eliminate gender leakage, while adversarial learning improves fairness mainly on original biographies and acts as a complementary strategy rather than a substitute for text-level debiasing.
CommentsAccepted for presentation at the TRUST-AI 2026 Workshop, held in conjunction with IJCAI/ECAI 2026, Bremen, Germany
