Recently, data-driven Artificial Intelligence (AI) techniques are increasingly being adopted for intelligent optimization and control of power and energy systems. The system status data (e.g., temperature) are collected and used to learn an AI system for intelligent control, management, and optimization towards energy efficiency. This research makes a special focus on innovating the safety and reliability assurance for AI systems applied in the context of energy sectors, the current research of which is still at a very early stage. For example, in this project, we intend to propose an automated framework for safety issue detection of AI-enabled energy systems (e.g., power plant, power grid). We would also explore and propose novel techniques to enhance the safety and reliability of next-generation AI-enabled energy systems.
AutoRepair: Automated Repair for AI-Enabled Cyber- Physical Systems under Safety-Critical Conditions
Article in professional or trade journals
ISR-LLM: Iterative Self-Refined Large Language Model for Long-Horizon Sequential Task Planning
Conference Proceedings
Mosaic: Model-based Safety Analysis Framework for AI-enabled Cyber-Physical Systems
Article in professional or trade journals
SIEGE: A Semantics-Guided Safety Enhancement Framework for AI-enabled Cyber-Physical Systems
Article in professional or trade journals
Self-Refined Large Language Model as Automated Reward Function Designer for Deep Reinforcement Learning in Robotics
Conference Proceedings
Towards Building AI-CPS with NVIDIA Isaac Sim: An Industrial Benchmark and Case Study for Robotics Manipulation
Conference Proceedings
When Cyber-Physical Systems Meet AI: A Benchmark, an Evaluation, and a Way Forward
Conference Proceedings