Knowledge Graph Representation and Reasoning Forum at DataFun Summit 2022

Background: Knowledge graphs summarize human prior knowledge and have significant academic value and broad application prospects. In the era of deep learning, representing entities and relations as vectors enables heterogeneous information fusion and powerful graph reasoning for intelligent decision‑making.

Event Details: On March 12, 9:00‑12:45, the DataFun Summit 2022 Knowledge Graph Online Summit hosts a Knowledge Representation and Reasoning Forum organized by senior researcher Wang Guangtao from JD Silicon Valley AI Research Institute.

Speaker Line‑up and Topics:

1. Zhang Yongqi (Senior Algorithm Researcher, Fourth Paradigm) – "Automated Knowledge Graph Representation Learning: From Triples to Subgraphs". The talk covers automated machine‑learning techniques for triple, path, and subgraph modeling, reviewing mainstream methods and Fourth Paradigm’s proprietary algorithms.

2. Ren Hongyu (Ph.D. student, Stanford) – "Multi‑step Reasoning on Knowledge Graphs". This presentation introduces a framework that embeds multi‑hop logical queries in latent space and designs neural logical operators, enabling first‑order logic reasoning on massive graphs and showcasing the SMORE codebase for graphs with over 90 million nodes.

3. Shang Chao (Researcher, JD Silicon Valley Research Institute) – "Temporal Knowledge Graph Question‑Answering System". The session explores how to incorporate time information into knowledge graphs, improve temporal QA sensitivity, and model time‑aware reasoning in natural language queries.

4. Xiao Nan (Algorithm Expert, JD Technology) – "Knowledge Reasoning in Financial Scenarios". The talk discusses applying causal knowledge in investment research and public opinion analysis, focusing on discovering and aligning causal relations.

5. Chen Fengjiao (Technical Expert, Meituan) – "Product Understanding with Meituan Brain". This presentation describes how Meituan’s knowledge base enables structured product understanding for retail, enhancing model generalization, efficiency, and practical problem solving.

Each speaker’s biography highlights their expertise in graph neural networks, knowledge graph construction, multi‑hop reasoning, and AI applications across finance, e‑commerce, and NLP.