Wang Haofen

Researcher Fellow, Doctoral Supervisor

College of Design and Innovation, Tongji University

Haofen Wang is a research fellow and doctoral supervisor at the College of Design and Innovation, Tongji University. His research focuses on knowledge graphs, natural language processing, retrieval-augmented generation, and knowledge-enhanced large language models. He leads the Knowledge Computing Lab at Tongji University, and is the leading founder of OpenKG, which is the largest Chinese knowledge graph community in the world. Currently, he holds important academic roles with the China Computer Federation (CCF), the Chinese Information Processing Society of China (CIPS), and the Shanghai Computer Federation, and has served as long-term PC member and reviewers for top journals and conferences on knowledge graphs and artificial intelligence, including AAAI, ACM MM, ICLR, IJCAI, IPM, ISWC, JWS, KBS, KDD, NeurIPS, TKDE, TNNLS, TOIS, WWW. Before joining Tongji, he served as CTO in leading AI companies, where he led a team in building an intelligent customer service robot based on knowledge graphs, that has served over 1 billion end-users. He has host several national natural science foundation projects , including one National Natural Science Key Project, 863 major projects, and relevant projects supported by the Ministry of Science and Technology. He has published over 100 high-level papers with over 9,300 citations, ,with a single paper cited over 4,000 times, five books on knowledge graphs, with the book "Knowledge Graph Methods, Practices, and Applications" winning the 2019 Publication Award of Professional Category of 2020 China Industry and Information Technology Publishing& Media Group (4/8000).

Education Experience:

Research and Academic Work Experience:

Research Interests

Academic Roles:

Publications

Published over 100 high-level papers in the Knowledge Graph, Natural Language Processing, and other AI fields, with more than 9,300 citations. [Selected Recent Papers]

Recent Research Grants

Open Source Projects

OpenKG [link]

OpenKG is an open knowledge graph initiative aimed at promoting the openness, interconnection, and crowdsourcing of knowledge graph data centered on Chinese, as well as the open-source development of knowledge graph tools, models, and platforms.

OpenKG

OpenSPG [link]

OpenSPG is a knowledge graph engine developed by Ant Group in collaboration with OpenKG, based on the SPG (Semantic-enhanced Programmable Graph) framework. It provides capabilities for domain model constrained knowledge modeling and facts and logic fused representation.

OpenSPG

KAG [link]

KAG is a professional domain knowledge-augmented service framework specifically designed for building domain-specific knowledge bases.

KAG

MemOS [link]

MemOS is an intelligent memory operating system that enhances the personalization of large models. Through two key mechanisms—memory tiering and multi-granular scheduling—it enables continuous evolution and personalized responses of the models.

MemOS

KaLM Embedding[link]

KaLM-Embedding, a multilingual embedding model, leverages high-quality training data and advanced techniques to achieve superior performance compared to other similarly sized models on the MTEB benchmark.

KALM

JoyAgent[link]

JoyAgent is the industry's first open-source, fully-featured, lightweight, and general-purpose multi-agent product.

JoyAgent

AI-Ceping [link]

AI-Ceping is a cutting-edge platform dedicated to the evaluation and advancement of LLMs. As a pioneer in the field, AI-Ceping offers a comprehensive suite of tools and services designed to test, improve, and showcase LLMs capabilities.

AI-Ceping

Awards

Selected Talks:

Practical Applications