Unified Shapley Framework to Explain Prediction Drift
This provides a systematic framework for understanding prediction drift in machine learning models, addressing a gap in research and practice for interpreting model behavior over data slices or temporal changes.
The authors tackled the problem of explaining prediction drift across data segments or time by proposing GroupShapley and GroupIG as axiomatically justified methods, unifying existing Shapley-based importance measures under a distributional comparison framework.
Predictions are the currency of a machine learning model, and to understand the model's behavior over segments of a dataset, or over time, is an important problem in machine learning research and practice. There currently is no systematic framework to understand this drift in prediction distributions over time or between two semantically meaningful slices of data, in terms of the input features and points. We propose GroupShapley and GroupIG (Integrated Gradients), as axiomatically justified methods to tackle this problem. In doing so, we re-frame all current feature/data importance measures based on the Shapley value as essentially problems of distributional comparisons, and unify them under a common umbrella. We axiomatize certain desirable properties of distributional difference, and study the implications of choosing them empirically.