Graphusion: Leveraging Large Language Models for Scientific Knowledge Graph Fusion and Construction in NLP Education

Rui Yang,Boming Yang,Sixun Ouyang,Tianwei She,Aosong Feng,Yuang Jiang,Freddy Lecue,Jinghui Lu,Irene Li

2024-07-15

Abstract:Knowledge graphs (KGs) are crucial in the field of artificial intelligence and are widely applied in downstream tasks, such as enhancing Question Answering (QA) systems. The construction of KGs typically requires significant effort from domain experts. Recently, Large Language Models (LLMs) have been used for knowledge graph construction (KGC), however, most existing approaches focus on a local perspective, extracting knowledge triplets from individual sentences or documents. In this work, we introduce Graphusion, a zero-shot KGC framework from free text. The core fusion module provides a global view of triplets, incorporating entity merging, conflict resolution, and novel triplet discovery. We showcase how Graphusion could be applied to the natural language processing (NLP) domain and validate it in the educational scenario. Specifically, we introduce TutorQA, a new expert-verified benchmark for graph reasoning and QA, comprising six tasks and a total of 1,200 QA pairs. Our evaluation demonstrates that Graphusion surpasses supervised baselines by up to 10% in accuracy on link prediction. Additionally, it achieves average scores of 2.92 and 2.37 out of 3 in human evaluations for concept entity extraction and relation recognition, respectively.

Computation and Language,Artificial Intelligence

What problem does this paper attempt to address?

The paper aims to address the challenges in constructing and utilizing Knowledge Graphs (KGs) for enhancing natural language processing (NLP) tasks, particularly in educational settings. The primary problems targeted by the paper include: 1. **Improving the Reliability and Interpretability of Content Generated by Large Language Models (LLMs):** Existing LLMs often generate content that lacks accuracy and interpretability. To address this, the authors propose leveraging KGs to enhance the performance of LLMs. 2. **Constructing Comprehensive and Accurate Knowledge Graphs:** Current approaches to Knowledge Graph Construction (KGC) often focus on extracting information from individual sentences or paragraphs, which limits the comprehensiveness and interconnectedness of the knowledge represented. The paper introduces a global perspective for KGC that aims to capture the broader context and relations that span beyond single documents. 3. **Applying Knowledge Graphs in Educational Scenarios:** There is a need for more effective question-answering (QA) systems in educational contexts that can provide relevant and accurate responses based on deep understanding of the concepts involved. The paper develops a framework and benchmark dataset to evaluate the use of KGs in educational QA. To solve these problems, the authors propose the Graphusion framework, which enables zero-shot KGC from free text. This framework includes a graph fusion module that extracts candidate triplets and per

Graphusion: Leveraging Large Language Models for Scientific Knowledge Graph Fusion and Construction in NLP Education

Graphusion: A RAG Framework for Knowledge Graph Construction with a Global Perspective

Graph Reasoning Transformers for Knowledge-Aware Question Answering

AGENTiGraph: An Interactive Knowledge Graph Platform for LLM-based Chatbots Utilizing Private Data

Leveraging Large Language Models for Concept Graph Recovery and Question Answering in NLP Education

KG-EGV: A Framework for Question Answering with Integrated Knowledge Graphs and Large Language Models

Efficient Knowledge Infusion via KG-LLM Alignment

Exploiting Large Language Models Capabilities for Question Answer-Driven Knowledge Graph Completion Across Static and Temporal Domains

Enhancing Large Language Models with Knowledge Graphs for Robust Question Answering

Can LLMs be Good Graph Judger for Knowledge Graph Construction?

Retrieval and Reasoning on KGs: Integrate Knowledge Graphs into Large Language Models for Complex Question Answering

SAC-KG: Exploiting Large Language Models as Skilled Automatic Constructors for Domain Knowledge Graphs

Multi-source Education Knowledge Graph Construction and Fusion for College Curricula

KICGPT: Large Language Model with Knowledge in Context for Knowledge Graph Completion

Collaborative Knowledge Graph Fusion by Exploiting the Open Corpus

JellyBell at TextGraphs-17 Shared Task: Fusing Large Language Models with External Knowledge for Enhanced Question Answering

LLM4EduKG: LLM for Automatic Construction of Educational Knowledge Graph

CogMG: Collaborative Augmentation Between Large Language Model and Knowledge Graph

LLMs for Knowledge Graph Construction and Reasoning: Recent Capabilities and Future Opportunities