When Machine Learning Gets Personal: Understanding Fairness of Personalized Models

2025-02-05

AI Summary

This paper addresses the tripartite trade-off among accuracy, interpretability, and fairness in personalized machine learning for high-stakes domains such as healthcare. Methodologically, we propose the first unified evaluation framework that: (1) establishes a joint metric for quantifying both predictive accuracy and explanation quality; (2) derives theoretical upper bounds on generalization error and fairness violation induced by personal attribute usage; and (3) integrates sensitivity analysis, feature attribution consistency metrics, and empirical validation on real-world clinical datasets. Key findings reveal that regression tasks tolerate personal attribute incorporation more than classification tasks, and accuracy gains do not necessarily improve explanation quality. To our knowledge, this is the first work to systematically characterize the interplay among accuracy, interpretability, and fairness in personalized modeling. The framework provides both theoretical foundations and practical guidelines for developing responsible, high-stakes personalized AI systems.

Technology Category

• Machine Learning: Privacy
• Philosophy and Ethics of AI: Accountability, Interpretability & Explainability
• Natural Language Processing: Interpretability, Analysis, and Evaluation of NLP Models

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Security and Privacy: Security and privacy of machine learning and AI applications

Abstract

Personalization in machine learning involves tailoring models to individual users by incorporating personal attributes such as demographic or medical data. While personalization can improve prediction accuracy, it may also amplify biases and reduce explainability. This work introduces a unified framework to evaluate the impact of personalization on both prediction accuracy and explanation quality across classification and regression tasks. We derive novel upper bounds for the number of personal attributes that can be used to reliably validate benefits of personalization. Our analysis uncovers key trade-offs. We show that regression models can potentially utilize more personal attributes than classification models. We also demonstrate that improvements in prediction accuracy due to personalization do not necessarily translate to enhanced explainability -- underpinning the importance to evaluate both metrics when personalizing machine learning models in critical settings such as healthcare. Validated with a real-world dataset, this framework offers practical guidance for balancing accuracy, fairness, and interpretability in personalized models.

Problem

• Evaluates personalization impact on accuracy, explainability.
• Derives bounds for personal attributes in validation.
• Balances fairness, accuracy, interpretability in models.

Innovation

• Unified framework for personalization evaluation
• Novel bounds on personal attribute usage
• Trade-offs between accuracy and explainability