Showcases

Ironlings: Heterogeneous Multi-Embodied Agent System

Embodied AI

Led by Prof. Lei Chen and Dr. Xujia Li, the Heterogeneous Embodied Intelligence Lab launched the Ironlings project to enable multi-agent collaboration for smart campus tasks. Key research areas include brain-level planning, cerebellum-level execution, world models, and embodied AI safety.

Industry Partners
Partner Logo Partner Logo

LingCopilot: An LLM-Empowered Quantitative Trading System

Fintech

Under the leadership of Prof. Lei Chen and Dr. Xujia Li, our Fintech team focuses on financial knowledge management, LLM-based quantitative investment, and risk control. Our flagship innovation, Lingcopliot, leverages AI to transform natural language concepts into backtestable, executable strategies throughout the trading lifecycle. The project was awarded the Gold Medal with Congratulations of the Jury at the 50th International Exhibition of Inventions of Geneva.

Industry Partners
Partner Logo Partner Logo Partner Logo
View details

Diga: Graph Diffusion for Explainable Anti‑Money Laundering

Banking

The Diga project is a flagship initiative under the Joint Laboratory of HKUST and WeBank, aiming to achieve high-recall, interpretable anti-money laundering monitoring using Graph Neural Networks. Ourteam introduces a novel guided graph diffusion method based on subgraph reconstruction to identify suspicious accounts in transaction networks at a scale of 200 million accounts.

Industry Partners
Partner Logo

Olli: Digital Twin Platform for Smart Vehicles Based on Industrial Internet of Things

Industrial Software

The Olli project focuses on developing an industrial software solution for the operation and maintenance of smart vehicles. Leveraging Digital Twin and Industrial Internet of Things technologies, the system provides real-time monitoring of autonomous vehicles, including driving metrics, location, and battery status. By integrating 3D design data, simulation tools, and AI models, it analyzes and predicts potential safety hazards, thereby achieving comprehensive prognostics and health management.

Industry Partners
Partner Logo Partner Logo

OPTIMA: Agent-Centric Database System

Database
OPTIMA

Traditional database systems are constrained by passive "query–execute–return" models. Our Agent DB Research Group now launches OPTIMA project, which advocates transitioning from static data repositories to a task-driven, agent-centric database architecture. Designed to support the perception, planning, and execution of database operations by autonomous agents, OPTIMA empowers proactive and efficient data interaction.

Industry Partners
Partner Logo Partner Logo