Welcome to our EDBT 2025 tutorial!

EDBT2025

Tutorial: Unifying Large Language Models and Knowledge Graphs for Question Answering: Recent Advances and Opportunities

Overview

Abstract

Large language models (LLMs) have demonstrated remarkable performance on several question-answering (QA) tasks because of their superior capabilities in natural language understanding and generation. On the other hand, due to poor reasoning capacity, outdated or lack of domain knowledge, expensive retraining costs, and limited context lengths of LLMs, LLM-based QA methods struggle with complex QA tasks such as multi-hop QAs and long-context QAs. Knowledge graphs (KGs) store graphbased structured knowledge which are effective for reasoning and interpretability since KGs accumulate and convey explicit relationships-based factual and domain-specific knowledge from the real world. To address the challenges and limitations of LLMbased QA, several research works that unify LLMs+KGs for QA have been proposed recently.

This tutorial aims to furnish an overview of the state-of-the-art advances in unifying LLMs with KGs for QA, by categorizing them into three groups according to the roles of KGs when unifying with LLMs. The metrics and benchmarking datasets for evaluating the methods of LLMs+KGs for QA are presented, and domain-specific industry applications and demonstrations will be showcased. The open challenges are summarized and the opportunities for data management are highlighted.

Speakers

Chuangtao Ma

Chuangtao Ma

Chuangtao Ma is a postdoctoral researcher at Aalborg University, Denmark. His research focuses on knowledge graphs, knowledge-augmented models, and their applications in data management. He is a member of the management committee of the COST action on the Global Network on Large-Scale, Cross-domain, and Multilingual Open Knowledge Graphs.

Yongrui Chen

Yongrui Chen

Yongrui Chen is a postdoctoral researcher at Southeast University, China. He specializes in incorporating structured and semi-structured knowledge into foundational LLMs, to improve their complex knowledge reasoning capability. He has presented numerous papers at prominent venues, including NeurIPS, TKDE, IJCAI, AAAI, ACL, ISWC, and NAACL.

Tianxing Wu

Tianxing Wu

Tianxing Wu is is an associate professor at Southeast University, China. He is one of the main contributors to build Chinese large-scale encyclopedic knowledge graph: Zhishi.me and schema knowledge graph: Linked Open Schema. He has published over 60 papers in top-tier venues, e.g., ICDE, SIGIR, ACL, AAAI, IJCAI, ECAI, ISWC, TKDE, TKDD, JWS, and WWWJ.

Arijit Khan

Arijit Khan

Arijit Khan is an IEEE senior member, an ACM distinguished speaker, and an associate professor at Aalborg University, Denmark. He published over 90 papers in premier data management and mining venues including ACM SIGMOD, VLDB, TKDE, ICDE, ICLR, SDM, USENIX ATC, EDBT, WWW, WSDM, CIKM, and TKDD. Arijit co-presented tutorials on emerging graph queries, applications, big graph systems, and graph machine learning at VLDB, DSAA, CIKM, and ICDE.

Haofen Wang

Haofen Wang

Haofen Wang is a Professor at Tongji University, China. He is one of the initiators of OpenKG, the world’s largest alliance for Chinese open knowledge graphs. He published over 100 high-level papers in the AI field, and developed the world’s first interactive virtual idol–``Amber Xuyan’’. Additionally, the intelligent customer service robots he built have served over 1 billion users.

Time Schedule

Time Topic Presenter
10:30 - 10:45 AM Motivation and Introduction Arijit Khan
10:45 - 11:10 AM Unifying LLMs with KGs for QA Chuangtao Ma
11:10 - 11:35 AM Advanced Topics on LLM+KG for QA Yongrui Chen
11:35 - 11:45 AM Break  
11:45 - 12:05 PM Evaluations and Applications Tianxing Wu
12:05 - 12:15 PM Opportunities for Data Management Arijit Khan
12:15 - 12:20 PM Future Directions Tianxing Wu
12:20 - 12:30 PM Q&A Session  

Materials and Slides

  • Materials: The covered papers, pointers to opensource codebase, datasets, and demonstrations are available on GitHub for public access.
  • Tutorial Paper: The tutorial paper is available on OpenProceedings.
  • Slides: PDF Slides