FW-Shapley: Real-Time Estimation of Weighted Shapley Values

2024-04-14

IEEE International Conference on Acoustics, Speech, and Signal Processing

AI Summary

To address the computational intractability of weighted Shapley values in high-dimensional settings—stemming from exponential complexity—this paper proposes the first scalable estimation framework grounded in weighted least squares (WLS) theory. Methodologically, it introduces: (1) a novel WLS-based analytical characterization of weighted Shapley values; (2) FW-Shapley, a label-free, theoretically guaranteed estimator integrating amortized learning and gradient-based optimization; and (3) enhanced robustness via Monte Carlo approximation and Shapley generalization theory. Experiments demonstrate that FW-Shapley improves Inclusion AUC by 27% in feature attribution tasks, achieves data valuation speedup of 14× over KNN Shapley while maintaining comparable accuracy, and significantly advances the practical deployment of weighted Shapley values for large-scale fair credit allocation.

Technology Category

• Machine Learning: Feature Construction/Reformulation
• Game Theory and Economic Paradigms: Fair Division
• Multiagent Systems: Mechanism Design

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Security and Privacy: Large-scale security measurements

Abstract

Fair credit assignment is essential in various machine learning (ML) applications, and Shapley values have emerged as a valuable tool for this purpose. However, in critical ML applications such as data valuation and feature attribution, the uniform weighting of Shapley values across subset cardinalities leads to unintuitive credit assignments. To address this, weighted Shapley values were proposed as a generalization, allowing different weights for subsets with different cardinalities. Despite their advantages, similar to Shapley values, Weighted Shapley values suffer from exponential compute costs, making them impractical for high-dimensional datasets. To tackle this issue, we present two key contributions. Firstly, we provide a weighted least squares characterization of weighted Shapley values. Next, using this characterization, we propose Fast Weighted Shapley (FW-Shapley), an amortized framework for efficiently computing weighted Shapley values using a learned estimator. We further show that our estimator's training procedure is theoretically valid even though we do not use ground truth Weighted Shapley values during training. On the feature attribution task, we outperform the learned estimator FastSHAP by 27% (on average) in terms of Inclusion AUC. For data valuation, we are much faster (14 times) while being comparable to the state-of-the-art KNN Shapley.

Problem

• Addresses unintuitive credit assignments in Shapley values.
• Reduces exponential compute costs of Weighted Shapley values.
• Proposes FW-Shapley for efficient computation in high-dimensional datasets.

Innovation

• Weighted least squares for Shapley values
• Amortized framework FW-Shapley for efficiency
• No ground truth required for training