发表机构
University of Amsterdam（阿姆斯特丹大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究提出一个结合强化学习多重实例学习、对抗性去偏和偏好条件超网络的多目标框架，用于学生风险预测，发现偏好条件无法稳定控制公平性与性能的权衡，需显式梯度平衡机制。
AI 中文摘要
从教育互动数据预测学生表现，需要既准确又足够透明以支持有意义干预的模型，而人口统计信息的引入增加了不公平预测的额外风险。本研究探讨了一个多目标框架，该框架结合了基于强化学习的多重实例学习（RL-MIL）、对抗性去偏和偏好条件超网络，用于学生风险预测。MIL将每个学生表示为一个弱标注互动实例的包，而RL代理选择信息丰富的实例用于下游分类。评估了两种超网络变体，以确定用户定义的偏好标量能否连续控制预测性能与均等几率之间的权衡。底层RL-MIL基线实现了强大的分类性能，但两种超网络扩展均表现出模式崩溃：改变偏好权重在预期的公平性-性能前沿上产生很少的系统性移动。该失败与目标主导、通过条件机制的弱梯度传播以及动态生成参数之间的相互作用有关。结果表明，公平目标可以整合到可解释的RL-MIL流程中，但仅偏好条件并不能保证可控的多目标行为。因此，稳健的公平RL-MIL需要明确的梯度平衡、目标分离和稳定性分析机制。
英文摘要
Predicting student performance from educational interaction data requires models that are both accurate and sufficiently transparent to support meaningful intervention, while demographic information introduces an additional risk of unfair predictions. This study investigates a multi-objective framework that combines reinforcement learning-based multiple instance learning (RL-MIL), adversarial debiasing, and preference-conditioned hypernetworks for student-at-risk prediction. MIL represents each student as a bag of weakly labeled interactions, while an RL agent selects informative instances for downstream classification. Two hypernetwork variants are evaluated to determine whether a user-defined preference scalar can continuously control the trade-off between predictive performance and Equalized Odds. The underlying RL-MIL baseline achieves strong classification performance, but both hypernetwork extensions exhibit mode collapse: changing the preference weight produces little systematic movement along the intended fairness-performance frontier. The failure is associated with objective dominance, weak gradient propagation through the conditioning mechanism, and interactions between dynamically generated parameters. The results show that fairness objectives can be incorporated into an interpretable RL-MIL pipeline, but preference conditioning alone does not guarantee controllable multi-objective behavior. Robust fair RL-MIL therefore requires explicit mechanisms for gradient balancing, objective separation, and stability analysis.
