Learning to Estimate Shapley Values with Vision Transformers
This work addresses the challenge of interpretability in vision transformers for researchers and practitioners, though it is incremental as it builds on existing Shapley value theory.
The paper tackled the problem of making Shapley values practical for explaining vision transformers by developing a learned explainer model, achieving more accurate explanations than baseline methods like attention rollout and GradCAM.
Transformers have become a default architecture in computer vision, but understanding what drives their predictions remains a challenging problem. Current explanation approaches rely on attention values or input gradients, but these provide a limited view of a model's dependencies. Shapley values offer a theoretically sound alternative, but their computational cost makes them impractical for large, high-dimensional models. In this work, we aim to make Shapley values practical for vision transformers (ViTs). To do so, we first leverage an attention masking approach to evaluate ViTs with partial information, and we then develop a procedure to generate Shapley value explanations via a separate, learned explainer model. Our experiments compare Shapley values to many baseline methods (e.g., attention rollout, GradCAM, LRP), and we find that our approach provides more accurate explanations than existing methods for ViTs.