QPaug: Question and Passage Augmentation for Open-Domain Question Answering of LLMs

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kim, Minsang, Park, Cheoneum, Baek, Seungjun
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913521130799104
author Kim, Minsang
Park, Cheoneum
Baek, Seungjun
author_facet Kim, Minsang
Park, Cheoneum
Baek, Seungjun
contents Retrieval-augmented generation (RAG) has received much attention for Open-domain question-answering (ODQA) tasks as a means to compensate for the parametric knowledge of large language models (LLMs). While previous approaches focused on processing retrieved passages to remove irrelevant context, they still rely heavily on the quality of retrieved passages which can degrade if the question is ambiguous or complex. In this paper, we propose a simple yet efficient method called question and passage augmentation (QPaug) via LLMs for open-domain QA. QPaug first decomposes the original questions into multiple-step sub-questions. By augmenting the original question with detailed sub-questions and planning, we are able to make the query more specific on what needs to be retrieved, improving the retrieval performance. In addition, to compensate for the case where the retrieved passages contain distracting information or divided opinions, we augment the retrieved passages with self-generated passages by LLMs to guide the answer extraction. Experimental results show that QPaug outperforms the previous state-of-the-art and achieves significant performance gain over existing RAG methods. The source code is available at \url{https://github.com/kmswin1/QPaug}.
format Preprint
id arxiv_https___arxiv_org_abs_2406_14277
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle QPaug: Question and Passage Augmentation for Open-Domain Question Answering of LLMs
Kim, Minsang
Park, Cheoneum
Baek, Seungjun
Computation and Language
Artificial Intelligence
Retrieval-augmented generation (RAG) has received much attention for Open-domain question-answering (ODQA) tasks as a means to compensate for the parametric knowledge of large language models (LLMs). While previous approaches focused on processing retrieved passages to remove irrelevant context, they still rely heavily on the quality of retrieved passages which can degrade if the question is ambiguous or complex. In this paper, we propose a simple yet efficient method called question and passage augmentation (QPaug) via LLMs for open-domain QA. QPaug first decomposes the original questions into multiple-step sub-questions. By augmenting the original question with detailed sub-questions and planning, we are able to make the query more specific on what needs to be retrieved, improving the retrieval performance. In addition, to compensate for the case where the retrieved passages contain distracting information or divided opinions, we augment the retrieved passages with self-generated passages by LLMs to guide the answer extraction. Experimental results show that QPaug outperforms the previous state-of-the-art and achieves significant performance gain over existing RAG methods. The source code is available at \url{https://github.com/kmswin1/QPaug}.
title QPaug: Question and Passage Augmentation for Open-Domain Question Answering of LLMs
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2406.14277