Precise Zero-Shot Pointwise Ranking with LLMs through Post-Aggregated Global Context Information

TL;DR

This paper introduces a novel zero-shot pointwise ranking method for large language models that incorporates global context through anchor-based contrastive scoring, significantly improving effectiveness while maintaining efficiency.

Contribution

The paper proposes GCCP and PAGC strategies that integrate global reference comparisons into pointwise ranking, enhancing performance without increasing computational costs.

Findings

Outperforms previous pointwise methods on TREC DL and BEIR benchmarks.

Achieves competitive results with more resource-intensive comparative methods.

Maintains efficiency comparable to traditional pointwise approaches.

Abstract

Recent advancements have successfully harnessed the power of Large Language Models (LLMs) for zero-shot document ranking, exploring a variety of prompting strategies. Comparative approaches like pairwise and listwise achieve high effectiveness but are computationally intensive and thus less practical for larger-scale applications. Scoring-based pointwise approaches exhibit superior efficiency by independently and simultaneously generating the relevance scores for each candidate document. However, this independence ignores critical comparative insights between documents, resulting in inconsistent scoring and suboptimal performance. In this paper, we aim to improve the effectiveness of pointwise methods while preserving their efficiency through two key innovations: (1) We propose a novel Global-Consistent Comparative Pointwise Ranking (GCCP) strategy that incorporates global reference comparisons between each candidate and an anchor document to generate contrastive relevance scores. We strategically design the anchor document as a query-focused summary of pseudo-relevant candidates, which serves as an effective reference point by capturing the global context for document comparison. (2) These contrastive relevance scores can be efficiently Post-Aggregated with existing pointwise methods, seamlessly integrating essential Global Context information in a training-free manner (PAGC). Extensive experiments on the TREC DL and BEIR benchmark demonstrate that our approach significantly outperforms previous pointwise methods while maintaining comparable efficiency. Our method also achieves competitive performance against comparative methods that require substantially more computational resources. More analyses further validate the efficacy of our anchor construction strategy.