Aligning Faithful Interpretations with their Social Attribution

TL;DR

This paper critiques the vagueness of faithfulness in model interpretations, proposing a formalization that aligns causal explanations with social attributions to improve interpretability.

Contribution

It introduces the concept of aligned faithfulness, formalizes causal and social attributions, and offers a causal reformulation to enhance faithful interpretability methods.

Findings

Identifies failures in existing highlight-based interpretations

Proposes a formal causal framework for faithful explanations

Demonstrates improved interpretability with contrastive explanations

Abstract

We find that the requirement of model interpretations to be faithful is vague and incomplete. With interpretation by textual highlights as a case-study, we present several failure cases. Borrowing concepts from social science, we identify that the problem is a misalignment between the causal chain of decisions (causal attribution) and the attribution of human behavior to the interpretation (social attribution). We re-formulate faithfulness as an accurate attribution of causality to the model, and introduce the concept of aligned faithfulness: faithful causal chains that are aligned with their expected social behavior. The two steps of causal attribution and social attribution together complete the process of explaining behavior. With this formalization, we characterize various failures of misaligned faithful highlight interpretations, and propose an alternative causal chain to remedy the issues. Finally, we implement highlight explanations of the proposed causal format using contrastive explanations.