IDEAS Lab
Welcome to the IDEAS Lab at Northwestern University! Our research interests include design automation for intelligent cyber-physical systems (CPS) and Internet-of-Things (IoT) applications, safe, robust and data-efficient machine learning for embodied AI systems, cyber-physical security, and energy-efficient CPS. Our recent projects have been focusing on the cross-layer design, verification, and adaptation of learning-enabled cyber-physical systems (LE-CPS), particularly addressing the safety, robustness, security, energy, and data challenges in utilizing machine learning for CPS and in operating CPS within dynamic and uncertain environment. We work on applications in the domains of connected and autonomous vehicles, smart buildings and infrastructures, robotics, advanced manufacturing, IoT, etc.
news
| May 2025 | [ICML 2025] Our co-authored paper “Directly Forecasting Belief for Reinforcement Learning with Delays” is accepted by the 42nd International Conference on Machine Learning (ICML). The work is in collaboration with University of Southampton and others. [ACL 2025] Paper “Can LLMs Understand Unvoiced Speech? Exploring EMG-to-Text Conversion with LLMs” is accepted by the 63rd Annual Meeting of the Association for Computational Linguistics (ACL) main conference. |
|---|---|
| Mar 2025 | [Scientific Reports] Paper “Efficient and Assured Reinforcement Learning-based Building HVAC Control with Heterogeneous Expert-guided Training” is published by Scientific Reports. [CVPR 2025] Paper “Split Adaptation for Pre-trained Vision Transformers” is accepted by the Conference on Computer Vision and Pattern Recognition (CVPR). |
| Jan 2025 | [ICLR 2025] Papers “On Large Language Model Continual Unlearning” and “Can We Trust Embodied Agents? Exploring Backdoor Attacks against Embodied LLM-Based Decision-Making Systems” are accepted by the 13th International Conference on Learning Representations (ICLR). |
| Sep 2024 | [Neurips 2024] Papers “Semantic Feature Learning for Universal Unsupervised Cross-Domain Retrieval” and “Variational Delayed Policy Optimization” are accepted by the 38th Annual Conference on Neural Information Processing Systems (Neurips). |
| Aug 2024 | [IROS 2024] Papers “Graph Neural Network-Based Multi-Agent Reinforcement Learning for Resilient Distributed Coordination of Multi-Robot Systems” and “Kinematics-Aware Trajectory Generation and Prediction with Latent Stochastic Differential Modeling” are accepted by the International Conference on Intelligent Robots and Systems (IROS). |