Soft Prompt Tuning for Augmenting Dense Retrieval with Large Language Models

TL;DR

This paper introduces a novel soft prompt tuning method to enhance dense retrieval models by generating high-quality weak queries with large language models, especially useful when domain-specific data is scarce.

Contribution

It proposes a new soft prompt tuning approach for augmenting dense retrieval, improving zero-shot and few-shot performance without relying on human-crafted prompts.

Findings

SPTAR outperforms BM25 and LLM-based augmentation methods.

Soft prompt tuning effectively generates high-quality weak queries.

The method improves dense retrieval in low-data scenarios.

Abstract

Dense retrieval (DR) converts queries and documents into dense embeddings and measures the similarity between queries and documents in vector space. One of the challenges in DR is the lack of domain-specific training data. While DR models can learn from large-scale public datasets like MS MARCO through transfer learning, evidence shows that not all DR models and domains can benefit from transfer learning equally. Recently, some researchers have resorted to large language models (LLMs) to improve the zero-shot and few-shot DR models. However, the hard prompts or human-written prompts utilized in these works cannot guarantee the good quality of generated weak queries. To tackle this, we propose soft prompt tuning for augmenting DR (SPTAR): For each task, we leverage soft prompt-tuning to optimize a task-specific soft prompt on limited ground truth data and then prompt the LLMs to tag unlabeled documents with weak queries, yielding enough weak document-query pairs to train task-specific dense retrievers. We design a filter to select high-quality example document-query pairs in the prompt to further improve the quality of weak tagged queries. To the best of our knowledge, there is no prior work utilizing soft prompt tuning to augment DR models. The experiments demonstrate that SPTAR outperforms the unsupervised baselines BM25 and the recently proposed LLMs-based augmentation method for DR.