Attributing Fair Decisions with Attention Interventions
This work addresses the need for auditable and fair AI decisions in domains like healthcare and parole systems, offering a method that combines fairness and explainability, though it is incremental as it builds on existing attention mechanisms.
The paper tackles the problem of ensuring fairness and providing explanations for AI decisions in high-stakes domains by proposing an attention-based model that identifies features affecting both performance and fairness through attention interventions, demonstrating its versatility on tabular and textual data with competitive results compared to baselines.
The widespread use of Artificial Intelligence (AI) in consequential domains, such as healthcare and parole decision-making systems, has drawn intense scrutiny on the fairness of these methods. However, ensuring fairness is often insufficient as the rationale for a contentious decision needs to be audited, understood, and defended. We propose that the attention mechanism can be used to ensure fair outcomes while simultaneously providing feature attributions to account for how a decision was made. Toward this goal, we design an attention-based model that can be leveraged as an attribution framework. It can identify features responsible for both performance and fairness of the model through attention interventions and attention weight manipulation. Using this attribution framework, we then design a post-processing bias mitigation strategy and compare it with a suite of baselines. We demonstrate the versatility of our approach by conducting experiments on two distinct data types, tabular and textual.