arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2310.09497cs.IRcs.AI
收藏

一种用于大型语言模型高效且高度有效的零样本排序的集合式方法

A Setwise Approach for Effective and Highly Efficient Zero-shot Ranking with Large Language Models

CSIRO(联邦科学与工业研究组织);Google Research(谷歌研究院);The University of Queensland(昆士兰大学)

Shengyao Zhuang, Honglei Zhuang, Bevan Koopman, Guido Zuccon

AI总结:

提出 Setwise prompting 零样本文档排序方法,通过减少 LLM 推理次数和 token 消耗,在保持高有效性的同时显著提升基于 LLM 的零样本排序效率。

摘要:

我们提出了一种基于大型语言模型(LLMs)的新型零样本文档排序方法:Setwise prompting 方法。该方法补充了现有的基于 LLM 的零样本排序提示方法:Pointwise、Pairwise 和 Listwise。通过在一致的实验框架内进行首次同类比较评估,并综合考虑模型规模、token 消耗和延迟等因素,我们发现现有方法在有效性和效率之间存在固有权衡。Pointwise 方法效率高但有效性差;相反,Pairwise 方法有效性优越但计算开销高。相比之下,我们的 Setwise 方法减少了排序过程中的 LLM 推理次数和提示 token 消耗量,从而显著提高了基于 LLM 的零样本排序效率,同时保持了高零样本排序有效性。我们的代码和结果已在 https://github.com/ielab/llm-rankers 公开。

英文摘要:

We propose a novel zero-shot document ranking approach based on Large Language Models (LLMs): the Setwise prompting approach. Our approach complements existing prompting approaches for LLM-based zero-shot ranking: Pointwise, Pairwise, and Listwise. Through the first-of-its-kind comparative evaluation within a consistent experimental framework and considering factors like model size, token consumption, latency, among others, we show that existing approaches are inherently characterised by trade-offs between effectiveness and efficiency. We find that while Pointwise approaches score high on efficiency, they suffer from poor effectiveness. Conversely, Pairwise approaches demonstrate superior effectiveness but incur high computational overhead. Our Setwise approach, instead, reduces the number of LLM inferences and the amount of prompt token consumption during the ranking procedure, compared to previous methods. This significantly improves the efficiency of LLM-based zero-shot ranking, while also retaining high zero-shot ranking effectiveness. We make our code and results publicly available at \url{https://github.com/ielab/llm-rankers}.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑