WeShap: Weak Supervision Source Evaluation with Shapley Values

2024-06-16

arXiv.org

AI Summary

Addressing the challenge of evaluating annotation quality from multiple sources in programmatic weak supervision (PWS), this paper introduces WeShap—a novel, Shapley-value-based metric specifically designed to quantify the contribution of individual labeling sources in weak supervision settings. By theoretically reformulating the Shapley value and incorporating dynamic programming optimization, WeShap reduces computational complexity to *O(m²)*, significantly enhancing scalability. It is interpretable, efficiently computable, and generalizable across diverse model architectures. WeShap enables labeling function diagnostics, correction of mislabeled instances, and end-to-end PWS pipeline optimization. Empirical evaluation across multiple benchmark datasets demonstrates that WeShap-guided annotation strategies improve downstream model accuracy by an average of 5.0 percentage points, thereby strengthening both the robustness and interpretability of PWS systems.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Multiagent Systems: Adversarial Agents

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

Efficient data annotation stands as a significant bottleneck in training contemporary machine learning models. The Programmatic Weak Supervision (PWS) pipeline presents a solution by utilizing multiple weak supervision sources to automatically label data, thereby expediting the annotation process. Given the varied contributions of these weak supervision sources to the accuracy of PWS, it is imperative to employ a robust and efficient metric for their evaluation. This is crucial not only for understanding the behavior and performance of the PWS pipeline but also for facilitating corrective measures. In our study, we introduce WeShap values as an evaluation metric, which quantifies the average contribution of weak supervision sources within a proxy PWS pipeline, leveraging the theoretical underpinnings of Shapley values. We demonstrate efficient computation of WeShap values using dynamic programming, achieving quadratic computational complexity relative to the number of weak supervision sources. Our experiments demonstrate the versatility of WeShap values across various applications, including the identification of beneficial or detrimental labeling functions, refinement of the PWS pipeline, and rectification of mislabeled data. Furthermore, WeShap values aid in comprehending the behavior of the PWS pipeline and scrutinizing specific instances of mislabeled data. Although initially derived from a specific proxy PWS pipeline, we empirically demonstrate the generalizability of WeShap values to other PWS pipeline configurations. Our findings indicate a noteworthy average improvement of 5.0 points in downstream model accuracy through the revision of the PWS pipeline compared to previous state-of-the-art methods, underscoring the efficacy of WeShap values in enhancing data quality for training machine learning models.

Problem

• Evaluates weak supervision sources using Shapley values
• Improves data annotation efficiency in machine learning
• Enhances downstream model accuracy by refining PWS pipeline

Innovation

• WeShap values evaluate weak supervision sources.
• Dynamic programming enables efficient WeShap computation.
• WeShap improves model accuracy by 5.0 points.