Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Towards Trustworthy Knowledge Graph Reasoning: An Uncertainty Aware Perspective

Bo Ni, Yu Wang, Lu Cheng, Erik Blasch, Tyler Derr

TL;DR

A new trustworthy KG-LLM framework, UAG (Uncertainty Aware Knowledge-Graph Reasoning), which incorporates uncertainty quantification into the KG-LLM framework, and design an uncertainty-aware multi-step reasoning framework that leverages conformal prediction to provide a theoretical guarantee on the prediction set.

Abstract

Recently, Knowledge Graphs (KGs) have been successfully coupled with Large Language Models (LLMs) to mitigate their hallucinations and enhance their reasoning capability, such as in KG-based retrieval-augmented frameworks. However, current KG-LLM frameworks lack rigorous uncertainty estimation, limiting their reliable deployment in high-stakes applications. Directly incorporating uncertainty quantification into KG-LLM frameworks presents challenges due to their complex architectures and the intricate interactions between the knowledge graph and language model components. To address this gap, we propose a new trustworthy KG-LLM framework, Uncertainty Aware Knowledge-Graph Reasoning (UAG), which incorporates uncertainty quantification into the KG-LLM framework. We design an uncertainty-aware multi-step reasoning framework that leverages conformal prediction to provide a theoretical guarantee on the prediction set. To manage the error rate of the multi-step process, we additionally introduce an error rate control module to adjust the error rate within the individual components. Extensive experiments show that our proposed UAG can achieve any pre-defined coverage rate while reducing the prediction set/interval size by 40% on average over the baselines.

Towards Trustworthy Knowledge Graph Reasoning: An Uncertainty Aware Perspective

TL;DR

A new trustworthy KG-LLM framework, UAG (Uncertainty Aware Knowledge-Graph Reasoning), which incorporates uncertainty quantification into the KG-LLM framework, and design an uncertainty-aware multi-step reasoning framework that leverages conformal prediction to provide a theoretical guarantee on the prediction set.

Abstract

Recently, Knowledge Graphs (KGs) have been successfully coupled with Large Language Models (LLMs) to mitigate their hallucinations and enhance their reasoning capability, such as in KG-based retrieval-augmented frameworks. However, current KG-LLM frameworks lack rigorous uncertainty estimation, limiting their reliable deployment in high-stakes applications. Directly incorporating uncertainty quantification into KG-LLM frameworks presents challenges due to their complex architectures and the intricate interactions between the knowledge graph and language model components. To address this gap, we propose a new trustworthy KG-LLM framework, Uncertainty Aware Knowledge-Graph Reasoning (UAG), which incorporates uncertainty quantification into the KG-LLM framework. We design an uncertainty-aware multi-step reasoning framework that leverages conformal prediction to provide a theoretical guarantee on the prediction set. To manage the error rate of the multi-step process, we additionally introduce an error rate control module to adjust the error rate within the individual components. Extensive experiments show that our proposed UAG can achieve any pre-defined coverage rate while reducing the prediction set/interval size by 40% on average over the baselines.

Paper Structure

This paper contains 48 sections, 3 theorems, 11 equations, 4 figures, 6 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Preliminiaries
  3. Conformal Prediction
  4. Learn Then Test (LTT)
  5. Problem Definition
  6. Method
  7. UQ-aware Candidate Retriever
  8. Retrieving Candidate Path
  9. Retrieving Candidate Neighbor
  10. UQ-aware Candidate Evaluator
  11. Global Error Rate Controller
  12. Experiment
  13. Datasets
  14. Evaluation Metrics
  15. Baselines
  16. ...and 33 more sections

Key Result

Theorem 1

Suppose $p_\lambda$ is super-uniform under $H_0^\lambda \forall \lambda \in \Lambda$. Let $\mathcal{T}$ be any FWER-controlling algorithm at level $\delta$. Then $\Lambda_{valid}$ satisfies Eq. eq:ltt.

Figures (4)

  • Figure 1: An illustration of uncertainty quantification in the context of knowledge graph question answering. By quantifying a confidence boundary (dashlines), UQ methods can ensure the accuracy when facing uncertainties.
  • Figure 2: An illustration of our UaG framework.
  • Figure 3: Uncertainty quantification results for UaG and the baselines.
  • Figure 4: Ablation studies where (a) shows the benefits of pre-training the sentence transformer used in graph traversal and (b) shows a comparison when using L1 similarity as opposed to cosine similarity within UaG.

Theorems & Definitions (6)

  • Theorem 1: Learn Then Test angelopoulos2021learn
  • Definition 1: Uncertainty quantification for multi-hop knowledge graph question answering
  • Remark 1
  • Theorem 2: Binomial tail bound p-values quach2023conformal
  • Theorem 3: Coverage Guarantee of UaG
  • proof