A Two-Stage Adaptation of Large Language Models for Text Ranking

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Zhang, Longhui, Zhang, Yanzhao, Long, Dingkun, Xie, Pengjun, Zhang, Meishan, Zhang, Min
Natura: Preprint
Pubblicazione: 2023
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866916268457590784
author Zhang, Longhui
Zhang, Yanzhao
Long, Dingkun
Xie, Pengjun
Zhang, Meishan
Zhang, Min
author_facet Zhang, Longhui
Zhang, Yanzhao
Long, Dingkun
Xie, Pengjun
Zhang, Meishan
Zhang, Min
contents Text ranking is a critical task in information retrieval. Recent advances in pre-trained language models (PLMs), especially large language models (LLMs), present new opportunities for applying them to text ranking. While supervised fine-tuning (SFT) with ranking data has been widely explored to better align PLMs with text ranking goals, previous studies have focused primarily on encoder-only and encoder-decoder PLMs. Research on leveraging decoder-only LLMs for text ranking remains scarce. An exception to this is RankLLaMA, which uses direct SFT to explore LLaMA's potential for text ranking. In this work, we propose a two-stage progressive paradigm to better adapt LLMs to text ranking. First, we conduct continual pre-training (CPT) of LLMs on a large weakly-supervised corpus. Second, we perform SFT, and propose an improved optimization strategy building upon RankLLaMA. Our experimental results on multiple benchmarks show that our approach outperforms previous methods in both in-domain and out-domain scenarios.
format Preprint
id arxiv_https___arxiv_org_abs_2311_16720
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle A Two-Stage Adaptation of Large Language Models for Text Ranking
Zhang, Longhui
Zhang, Yanzhao
Long, Dingkun
Xie, Pengjun
Zhang, Meishan
Zhang, Min
Information Retrieval
Text ranking is a critical task in information retrieval. Recent advances in pre-trained language models (PLMs), especially large language models (LLMs), present new opportunities for applying them to text ranking. While supervised fine-tuning (SFT) with ranking data has been widely explored to better align PLMs with text ranking goals, previous studies have focused primarily on encoder-only and encoder-decoder PLMs. Research on leveraging decoder-only LLMs for text ranking remains scarce. An exception to this is RankLLaMA, which uses direct SFT to explore LLaMA's potential for text ranking. In this work, we propose a two-stage progressive paradigm to better adapt LLMs to text ranking. First, we conduct continual pre-training (CPT) of LLMs on a large weakly-supervised corpus. Second, we perform SFT, and propose an improved optimization strategy building upon RankLLaMA. Our experimental results on multiple benchmarks show that our approach outperforms previous methods in both in-domain and out-domain scenarios.
title A Two-Stage Adaptation of Large Language Models for Text Ranking
topic Information Retrieval
url https://arxiv.org/abs/2311.16720