Few-shot Prompting for Pairwise Ranking: An Effective Non-Parametric Retrieval Model

TL;DR

This paper introduces a pairwise few-shot prompting method for ranking that enhances zero-shot performance by leveraging similar query preferences, achieving results close to supervised models without complex training.

Contribution

The paper proposes a novel pairwise few-shot prompting approach that improves zero-shot ranking performance by incorporating examples from similar queries, reducing the need for supervised training.

Findings

Consistent improvements over zero-shot baseline on TREC DL and BEIR benchmarks.

Achieves near-supervised performance without complex training pipelines.

Effective non-parametric retrieval model based on few-shot prompting.

Abstract

A supervised ranking model, despite its advantage of being effective, usually involves complex processing - typically multiple stages of task-specific pre-training and fine-tuning. This has motivated researchers to explore simpler pipelines leveraging large language models (LLMs) that are capable of working in a zero-shot manner. However, since zero-shot inference does not make use of a training set of pairs of queries and their relevant documents, its performance is mostly worse than that of supervised models, which are trained on such example pairs. Motivated by the existing findings that training examples generally improve zero-shot performance, in our work, we explore if this also applies to ranking models. More specifically, given a query and a pair of documents, the preference prediction task is improved by augmenting examples of preferences for similar queries from a training set. Our proposed pairwise few-shot ranker demonstrates consistent improvements over the zero-shot baseline on both in-domain (TREC DL) and out-domain (BEIR subset) retrieval benchmarks. Our method also achieves a close performance to that of a supervised model without requiring any complex training pipeline.