Saved in:
Bibliographic Details
Main Authors: Wang, Zhengren, Yu, Qinhan, Wei, Shida, Li, Zhiyu, Xiong, Feiyu, Wang, Xiaoxing, Niu, Simin, Liang, Hao, Zhang, Wentao
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2409.20434
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913922403008512
author Wang, Zhengren
Yu, Qinhan
Wei, Shida
Li, Zhiyu
Xiong, Feiyu
Wang, Xiaoxing
Niu, Simin
Liang, Hao
Zhang, Wentao
author_facet Wang, Zhengren
Yu, Qinhan
Wei, Shida
Li, Zhiyu
Xiong, Feiyu
Wang, Xiaoxing
Niu, Simin
Liang, Hao
Zhang, Wentao
contents Modern QA systems entail retrieval-augmented generation (RAG) for accurate and trustworthy responses. However, the inherent gap between user queries and relevant documents hinders precise matching. We introduce QAEncoder, a training-free approach to bridge this gap. Specifically, QAEncoder estimates the expectation of potential queries in the embedding space as a robust surrogate for the document embedding, and attaches document fingerprints to effectively distinguish these embeddings. Extensive experiments across diverse datasets, languages, and embedding models confirmed QAEncoder's alignment capability, which offers a simple-yet-effective solution with zero additional index storage, retrieval latency, training costs, or catastrophic forgetting and hallucination issues. The repository is publicly available at https://github.com/IAAR-Shanghai/QAEncoder.
format Preprint
id arxiv_https___arxiv_org_abs_2409_20434
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle QAEncoder: Towards Aligned Representation Learning in Question Answering Systems
Wang, Zhengren
Yu, Qinhan
Wei, Shida
Li, Zhiyu
Xiong, Feiyu
Wang, Xiaoxing
Niu, Simin
Liang, Hao
Zhang, Wentao
Computation and Language
Modern QA systems entail retrieval-augmented generation (RAG) for accurate and trustworthy responses. However, the inherent gap between user queries and relevant documents hinders precise matching. We introduce QAEncoder, a training-free approach to bridge this gap. Specifically, QAEncoder estimates the expectation of potential queries in the embedding space as a robust surrogate for the document embedding, and attaches document fingerprints to effectively distinguish these embeddings. Extensive experiments across diverse datasets, languages, and embedding models confirmed QAEncoder's alignment capability, which offers a simple-yet-effective solution with zero additional index storage, retrieval latency, training costs, or catastrophic forgetting and hallucination issues. The repository is publicly available at https://github.com/IAAR-Shanghai/QAEncoder.
title QAEncoder: Towards Aligned Representation Learning in Question Answering Systems
topic Computation and Language
url https://arxiv.org/abs/2409.20434