Adaptive Bounded Exploration and Intermediate Actions for Data Debiasing

2025-04-10

AI Summary

This paper addresses classification under biased training data featuring cost-sensitive and truncated feedback. We propose a sequential debiasing framework that integrates adaptive exploration boundary control with low-cost intermediate action modeling to safely exploit noisy labels for accelerated fairness improvement. The method unifies sequential decision-making, causal inference, and constrained online learning, and we theoretically establish its efficacy in mitigating data bias under specific distributional assumptions. Experiments on multiple synthetic and real-world datasets demonstrate that our approach significantly outperforms baselines: classification accuracy and group fairness (measured by demographic parity and equalized odds) improve markedly, exploration cost decreases by 37%, and debiasing convergence accelerates by 2.1×. Our core contribution is the first introduction of an adaptive boundary-controlled, intermediate-action-driven sequential debiasing paradigm—establishing a novel principled approach to fair learning under costly, partial-feedback settings.

Technology Category

• Machine Learning: Causal Learning
• Search and Optimization: Learning to Search
• Reasoning under Uncertainty: Sequential Decision Making

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Economics, Online Markets and Human Computation: Fairness, privacy, and diversity in economic environments

Abstract

The performance of algorithmic decision rules is largely dependent on the quality of training datasets available to them. Biases in these datasets can raise economic and ethical concerns due to the resulting algorithms' disparate treatment of different groups. In this paper, we propose algorithms for sequentially debiasing the training dataset through adaptive and bounded exploration in a classification problem with costly and censored feedback. Our proposed algorithms balance between the ultimate goal of mitigating the impacts of data biases -- which will in turn lead to more accurate and fairer decisions, and the exploration risks incurred to achieve this goal. Specifically, we propose adaptive bounds to limit the region of exploration, and leverage intermediate actions which provide noisy label information at a lower cost. We analytically show that such exploration can help debias data in certain distributions, investigate how {algorithmic fairness interventions} can work in conjunction with our proposed algorithms, and validate the performance of these algorithms through numerical experiments on synthetic and real-world data.

Problem

• Mitigate biases in training datasets for fairer algorithmic decisions
• Balance debiasing goals with exploration risks in classification
• Use adaptive bounds and intermediate actions to reduce bias

Innovation

• Adaptive bounds limit exploration region
• Leverage intermediate actions for cost efficiency
• Balance debiasing and exploration risks