Lei Chen

AI for Finance Group

Overview

Financial knowledge and intelligent decision support

The AI for Finance Group studies how financial knowledge, graph learning and large language models can support financial analysis and decision-making. Its research connects dynamic financial knowledge base construction with stock prediction, quantitative investment, financial question answering and model explainability.

The group develops methods for extracting, integrating and updating financial knowledge, learning from evolving data, and applying that knowledge in interpretable financial AI systems.

Research areas

  • Financial knowledge base construction. Extracting knowledge, linking multimodal entities and updating financial facts to build dynamic knowledge bases. [HIT, Triple-d, T-FinKB, OpenMEL]
  • Temporal graph learning. Learning evolving graph representations while preserving earlier knowledge and adapting to new classes. [TE-DyGE, N-ForGOT]
  • Stock prediction and quantitative investment. Combining external knowledge and market signals for stock prediction, and using LLM agents to generate, evaluate and adapt trading strategies. [Survey, STEPS-MATA, LLM strategy finding]
  • Financial question answering. Retrieval-augmented generation and structured, multi-view retrieval for long-context and knowledge-based question answering. [RAG-Sketch, Multi-view VQA]
  • Model explainability. Heterogeneity-agnostic, multi-level explanations and data-aware global explanations for graph neural networks. [HENCE-X, On Data-Aware Global Explainability]