Incorporating Relevance Feedback for Information-Seeking Retrieval using Few-Shot Document Re-Ranking

TL;DR

This paper investigates integrating relevance feedback into neural document re-ranking models using few-shot and parameter-efficient learning, significantly improving retrieval performance in information-seeking scenarios.

Contribution

It introduces a kNN re-ranking approach and a meta-trained Cross-Encoder model that incorporate user relevance feedback directly into neural re-ranking, enhancing retrieval accuracy.

Findings

Relevance feedback integration improves neural re-ranking performance.

Fusing lexical and neural re-ranking outperforms other methods by 5.2 nDCG@20.

Meta-learning enhances Cross-Encoder adaptation to feedback.

Abstract

Pairing a lexical retriever with a neural re-ranking model has set state-of-the-art performance on large-scale information retrieval datasets. This pipeline covers scenarios like question answering or navigational queries, however, for information-seeking scenarios, users often provide information on whether a document is relevant to their query in form of clicks or explicit feedback. Therefore, in this work, we explore how relevance feedback can be directly integrated into neural re-ranking models by adopting few-shot and parameter-efficient learning techniques. Specifically, we introduce a kNN approach that re-ranks documents based on their similarity with the query and the documents the user considers relevant. Further, we explore Cross-Encoder models that we pre-train using meta-learning and subsequently fine-tune for each query, training only on the feedback documents. To evaluate our different integration strategies, we transform four existing information retrieval datasets into the relevance feedback scenario. Extensive experiments demonstrate that integrating relevance feedback directly in neural re-ranking models improves their performance, and fusing lexical ranking with our best performing neural re-ranker outperforms all other methods by 5.2 nDCG@20.