Adaptive Contrastive Decoding in Retrieval-Augmented Generation for Handling Noisy Contexts

TL;DR

This paper introduces adaptive contrastive decoding (ACD), a novel method that enhances retrieval-augmented generation by effectively handling noisy contexts, leading to improved robustness and accuracy in open-domain question answering tasks.

Contribution

The paper proposes ACD, an adaptive decoding approach that improves robustness of retrieval-augmented generation against noisy contexts, extending prior contrastive decoding methods.

Findings

ACD outperforms baseline methods in open-domain question answering.

ACD maintains accuracy even with noisy external contexts.

The approach enhances the robustness of retrieval-augmented generation models.

Abstract

When using large language models (LLMs) in knowledge-intensive tasks, such as open-domain question answering, external context can bridge the gap between external knowledge and the LLMs' parametric knowledge. Recent research has been developed to amplify contextual knowledge over the parametric knowledge of LLMs with contrastive decoding approaches. While these approaches could yield truthful responses when relevant context is provided, they are prone to vulnerabilities when faced with noisy contexts. We extend the scope of previous studies to encompass noisy contexts and propose adaptive contrastive decoding (ACD) to leverage contextual influence effectively. ACD demonstrates improvements in open-domain question answering tasks compared to baselines, especially in robustness by remaining undistracted by noisy contexts in retrieval-augmented generation.