InstUPR : Instruction-based Unsupervised Passage Reranking with Large Language Models
This addresses the problem of improving retrieval accuracy in information retrieval systems without labeled data, representing an incremental advance in unsupervised reranking techniques.
The paper tackles unsupervised passage reranking by introducing InstUPR, a method that uses instruction-tuned large language models without fine-tuning, achieving superior performance over unsupervised baselines and an instruction-tuned reranker on the BEIR benchmark.
This paper introduces InstUPR, an unsupervised passage reranking method based on large language models (LLMs). Different from existing approaches that rely on extensive training with query-document pairs or retrieval-specific instructions, our method leverages the instruction-following capabilities of instruction-tuned LLMs for passage reranking without any additional fine-tuning. To achieve this, we introduce a soft score aggregation technique and employ pairwise reranking for unsupervised passage reranking. Experiments on the BEIR benchmark demonstrate that InstUPR outperforms unsupervised baselines as well as an instruction-tuned reranker, highlighting its effectiveness and superiority. Source code to reproduce all experiments is open-sourced at https://github.com/MiuLab/InstUPR