Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Do You Know What You Are Talking About? Characterizing Query-Knowledge Relevance For Reliable Retrieval Augmented Generation

Zhuohang Li, Jiaxin Zhang, Chao Yan, Kamalika Das, Sricharan Kumar, Murat Kantarcioglu, Bradley A. Malin

TL;DR

A statistical framework is established that assesses how well a query can be answered by an RAG system by capturing the relevance of knowledge and an online testing procedure that employs goodness-of-fit (GoF) tests to inspect the relevance of each user query to detect out-of-knowledge queries with low knowledge relevance is introduced.

Abstract

Language models (LMs) are known to suffer from hallucinations and misinformation. Retrieval augmented generation (RAG) that retrieves verifiable information from an external knowledge corpus to complement the parametric knowledge in LMs provides a tangible solution to these problems. However, the generation quality of RAG is highly dependent on the relevance between a user's query and the retrieved documents. Inaccurate responses may be generated when the query is outside of the scope of knowledge represented in the external knowledge corpus or if the information in the corpus is out-of-date. In this work, we establish a statistical framework that assesses how well a query can be answered by an RAG system by capturing the relevance of knowledge. We introduce an online testing procedure that employs goodness-of-fit (GoF) tests to inspect the relevance of each user query to detect out-of-knowledge queries with low knowledge relevance. Additionally, we develop an offline testing framework that examines a collection of user queries, aiming to detect significant shifts in the query distribution which indicates the knowledge corpus is no longer sufficiently capable of supporting the interests of the users. We demonstrate the capabilities of these strategies through a systematic evaluation on eight question-answering (QA) datasets, the results of which indicate that the new testing framework is an efficient solution to enhance the reliability of existing RAG systems.

Do You Know What You Are Talking About? Characterizing Query-Knowledge Relevance For Reliable Retrieval Augmented Generation

TL;DR

A statistical framework is established that assesses how well a query can be answered by an RAG system by capturing the relevance of knowledge and an online testing procedure that employs goodness-of-fit (GoF) tests to inspect the relevance of each user query to detect out-of-knowledge queries with low knowledge relevance is introduced.

Abstract

Language models (LMs) are known to suffer from hallucinations and misinformation. Retrieval augmented generation (RAG) that retrieves verifiable information from an external knowledge corpus to complement the parametric knowledge in LMs provides a tangible solution to these problems. However, the generation quality of RAG is highly dependent on the relevance between a user's query and the retrieved documents. Inaccurate responses may be generated when the query is outside of the scope of knowledge represented in the external knowledge corpus or if the information in the corpus is out-of-date. In this work, we establish a statistical framework that assesses how well a query can be answered by an RAG system by capturing the relevance of knowledge. We introduce an online testing procedure that employs goodness-of-fit (GoF) tests to inspect the relevance of each user query to detect out-of-knowledge queries with low knowledge relevance. Additionally, we develop an offline testing framework that examines a collection of user queries, aiming to detect significant shifts in the query distribution which indicates the knowledge corpus is no longer sufficiently capable of supporting the interests of the users. We demonstrate the capabilities of these strategies through a systematic evaluation on eight question-answering (QA) datasets, the results of which indicate that the new testing framework is an efficient solution to enhance the reliability of existing RAG systems.

Paper Structure

This paper contains 42 sections, 4 equations, 6 figures, 16 tables.

Table of Contents

  1. Introduction
  2. Background
  3. Retrieval Augmented Generation (RAG).
  4. Embedding Model.
  5. Goodness-of-Fit (GoF) Test.
  6. A Statistical Characterization of Query-Knowledge Relevance
  7. Problem Definition
  8. Online Testing Procedure For Identifying Single Out-of-Knowledge Query
  9. Test Statistics.
  10. Meta-analytic Testing.
  11. Synthesizing In-Knowledge Queries.
  12. Offline Testing Procedure For Detecting Query Distribution Shift
  13. Experiments
  14. Experimental Setup
  15. Datasets and Corpora.
  16. ...and 27 more sections

Figures (6)

  • Figure 1: Overview of the hypothesis-testing framework for assessing query-knowledge relevance in RAG.
  • Figure 2: Illustration of critical values estimated using true in-knowledge and synthetic queries and histograms on the Textbooks corpus with energy score as test statistic.
  • Figure 3: Comparison of six different embedding models.
  • Figure 4: AUROC for PubMedQA as a function of $k$.
  • Figure 5: Offline testing results with the Textbooks corpus.
  • ...and 1 more figures

Theorems & Definitions (4)

  • Definition 3.1: Query-Knowledge Relevance
  • Definition 3.2: In-Knowledge Query
  • Definition 3.3: Out-of-Knowledge Query
  • Definition 3.4: GoF Test for Query Relevance