Interpretability Analysis of Domain Adapted Dense Retrievers

TL;DR

This paper introduces an interpretability method using integrated gradients to analyze how domain adaptation affects dense retrievers' focus on domain-specific terminology, enhancing understanding of their behavior across different domains.

Contribution

It develops a novel explainability framework for dense retrievers using integrated gradients and a new baseline, revealing how domain adaptation shifts model focus to in-domain terms.

Findings

Domain-adapted models focus more on in-domain terminology.

Integrated gradients effectively explain dense retriever behavior.

The method provides both instance-based and ranking-based explanations.

Abstract

Dense retrievers have demonstrated significant potential for neural information retrieval; however, they exhibit a lack of robustness to domain shifts, thereby limiting their efficacy in zero-shot settings across diverse domains. Previous research has investigated unsupervised domain adaptation techniques to adapt dense retrievers to target domains. However, these studies have not focused on explainability analysis to understand how such adaptations alter the model's behavior. In this paper, we propose utilizing the integrated gradients framework to develop an interpretability method that provides both instance-based and ranking-based explanations for dense retrievers. To generate these explanations, we introduce a novel baseline that reveals both query and document attributions. This method is used to analyze the effects of domain adaptation on input attributions for query and document tokens across two datasets: the financial question answering dataset (FIQA) and the biomedical information retrieval dataset (TREC-COVID). Our visualizations reveal that domain-adapted models focus more on in-domain terminology compared to non-adapted models, exemplified by terms such as "hedge," "gold," "corona," and "disease." This research addresses how unsupervised domain adaptation techniques influence the behavior of dense retrievers when adapted to new domains. Additionally, we demonstrate that integrated gradients are a viable choice for explaining and analyzing the internal mechanisms of these opaque neural models.