Papers
arxiv:2411.07021

Invar-RAG: Invariant LLM-aligned Retrieval for Better Generation

Published on Nov 11, 2024
Authors:
,
,
,
,

Abstract

Invar-RAG, a two-stage fine-tuning architecture, enhances retrieval and generation in RAG systems using LoRA and invariance loss, outperforming existing methods in open-domain question answering.

AI-generated summary

Retrieval-augmented generation (RAG) has shown impressive capability in providing reliable answer predictions and addressing hallucination problems. A typical RAG implementation uses powerful retrieval models to extract external information and large language models (LLMs) to generate answers. In contrast, recent LLM-based retrieval has gained attention for its substantial improvements in information retrieval (IR) due to the LLMs' semantic understanding capability. However, directly applying LLM to RAG systems presents challenges. This may cause feature locality problems as massive parametric knowledge can hinder effective usage of global information across the corpus; for example, an LLM-based retriever often inputs document summaries instead of full documents. Moreover, various pre-trained tasks in LLMs introduce variance, further weakening performance as a retriever. To address these issues, we propose a novel two-stage fine-tuning architecture called Invar-RAG. In the retrieval stage, an LLM-based retriever is constructed by integrating LoRA-based representation learning to tackle feature locality issues. To enhance retrieval performance, we develop two patterns (invariant and variant patterns) and an invariance loss to reduce LLM variance. In the generation stage, a refined fine-tuning method is employed to improve LLM accuracy in generating answers based on retrieved information. Experimental results show that Invar-RAG significantly outperforms existing baselines across three open-domain question answering (ODQA) datasets. Code is available in the Supplementary Material for reproducibility.

Community

Sign up or log in to comment

Models citing this paper 0

No model linking this paper

Cite arxiv.org/abs/2411.07021 in a model README.md to link it from this page.

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2411.07021 in a dataset README.md to link it from this page.

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2411.07021 in a Space README.md to link it from this page.

Collections including this paper 0

No Collection including this paper

Add this paper to a collection to link it from this page.