Lin ZHAO
Welcome!
I am currently an assistant professor in the Department of Electrical and Computer Engineering at the National University of Singapore.
I received my PhD degree in ECE and MSc in Mathematics from The Ohio State University, USA, in 2017, supervised by Prof. Wei Zhang and Prof. Ghaith A. Hiary, respectively. Before that, I received my B.S. and M.S. in Control Science and Engineering from Harbin Institute of Technology, China, in 2010 and 2012, respectively, supervised by Prof. Huijun Gao. Before joining NUS, I was a research scientist at Aptiv Pittsburgh Technology Center (now Motional) working on autonomous driving.
My research intersects control theory, optimization, and machine learning, with applications to unmanned aerial vehicles (UAV). One recent focus is on design and analysis of reinforcement learning algorithms, to boost training efficiencies and control performance with theoretical guarantees. On the UAV side, our group develop AI-assisted planning and control algorithms to enable more intelligent and robust autonomous operations and multi-agent collaborations.
News:
2024/04: Our paper "Global Optimality of Single-Timescale Actor-Critic under Continuous State-Action Space: a Study on Linear Quadratic Regulator" has been accepted by IJCAI 2024. Congrats to Xuyang!
2024/04: I am invited to give a talk in the workshop on Frontiers of optimization for robotics at the RSS 2024 in Delft on July 15.
2024/02: Wang Fenglan (PhD from Northeastern University, China) and He Lei (PhD from NWPU) joined our group! Welcome!
2024/01: Our paper "Trust-Region Neural Moving Horizon Estimation for Robots" has been accepted by ICRA 2024. Congrats to Bingheng and Xuyang!
2023/11: I am honored to serve on the Editorial Board of Unmanned Systems.
2023/10: Our paper "Neural Moving Horizon Estimation for Robust Flight Control" has been accepted by IEEE Transactions on Robotics. Congrats to Bingheng, Zhengtian, and Shupeng!
2023/09: Our paper "Finite-time analysis of single-timescale actor-critic" has been accepted by NeurIPS 2023 (acceptance rate 26.1%). Congrats to Chen Xuyang!
2023/08: Our project on fast adaptive meta-learning control of aerial robotic systems has been awarded by the Singapore Ministry of Education Academic Research Fund (AcRF) Tier 2 Grant!
2023/08: Liu Zhiyuan (BSc, MSc from HIT) and Tan Yan Rui (BSc from NUS) joined our group as PhD students. Welcome!
2023/07: My Co-PI'ed project "Development of Stable Robust and Secure Intelligent Systems for Autonomous Vehicles" (PI: Prof. Shuzhi Sam Ge) won the AI Singapore Grand Challenge Award in Robust AI.
2023/05: Our paper On the Optimization Landscape of Dynamic Output Feedback Linear Quadratic Control has been accepted by IEEE Transaction on Automatic Control as a full paper. Congrats to Jingliang!
2023/01: Our paper "Learning Agile Flight Maneuvers: Deep SE(3) Motion Planning and Control for Quadrotors" has been accepted by ICRA 2023. This work is spearheaded by Yixiao Wang (AY 2022 MSc EE graduate, now at Huawei) and Ph.D. student Bingheng Wang, with valuable contributions from Shenning Zhang (AY 2022 MSc EE graduate) and Sia Han Wei (AY 2021 MSc EE graduate, now at ST Engineering). Kudos to my talented students!
2022/11: Our paper "Global Convergence of Two-timescale Actor-Critic for Solving Linear Quadratic Regulator" has been accepted by the 37th AAAI Conference on Artificial Intelligence (AAAI-23) (acceptance rate 19.6%). This work is spearheaded by Xuyang Chen. Congrats to Xuyang!
2022/10: New manuscript on: "Finite-time analysis of single-timescale actor-critic" in general MDP setting
2022/10: I gave a talk on "Meta-Learning for Robot Control via Bilevel Optimization" at the Pre-conference workshop of The 4th International Forum on Frontiers of Data-driven Optimization of Complex Systems (FDOCS2022).
2022/10: IEEE CDC 2023 Website is live: https://cdc2023.ieeecss.org
2022/09: New manuscript on learning-based planning and control: "Learning Agile Flight Maneuvers: Deep SE(3) Motion Planning and Control for Quadrotors"
2022/08: New manuscript on analyzing Actor-Critic: "Global Convergence of Two-timescale Actor-Critic for Solving Linear Quadratic Regulator"
2022/08: I am honored to serve on the Young Editorial Board of the Springer Journal of Systems Science and Complexity.
2022/07: New manuscript on AI-powered estimation for robust control: "Neural Moving Horizon Estimation for Robust Flight Control"
2022/07: Our paper "On the Optimization Landscape of Dynamic Output Feedback: A Case Study for Linear Quadratic Regulator" has been accepted by the 61st IEEE Conference on Decision and Control (CDC 2022). Congrats to Jingliang! [arXiv]
2022/05: Our paper "Deterministic Policy Gradient: Convergence Analysis" has been accepted by 38th Conference on Uncertainty in Artificial Intelligence (UAI 2022).
2022/03: Zhao Kai (Research fellow, University of Macau) joined our group. Welcome!
2022/02: Our paper "Optimization Landscape of Gradient Descent for Discrete-time Static Output Feedback" has been accepted by the 2022 American Control Conference (ACC 2022). Congrats to Jingliang!
2022/01: I am honored to serve as the publicity co-chair of the 62nd IEEE Conference on Decision and Control (CDC 2023) that will be held in Marina Bay Sands, Singapore, Dec. 13-15, 2023. Welcome to Singapore!
2021/11: I am honored to serve as the program co-chair of the 17th IEEE International Conference on Control & Automation (IEEE ICCA 2022), planned to be held in Naples, Italy, June 27-30, 2022. Welcome to submit to our conference!
2021/09: Our paper "Faster Non-asymptotic Convergence for Double Q-learning" has been accepted by 35th Conference on Neural Information Processing Systems (NeurIPS 2021). [Slides, video]
2021/07: Our paper "Differentiable Moving Horizon Estimation for Robust Flight Control" has been accepted by the 60th IEEE Conference on Decision and Control (CDC 2021). [Video]
2021/07: Duan Jingliang (Research fellow, PhD from Tsinghua University) and Chen Xuyang (PhD student, BSc from BUAA) joined our group. Welcome!
2021/05: Our paper "Finite-Time Theory for Momentum Q-learning" has been accepted by 37th Conference on Uncertainty in Artificial Intelligence (UAI 2021).
2021/01: Ma Zhengtian joined our group as a PhD student. Welcome!
2020/09: Our paper "Finite-Time Analysis for Double Q-learning" has been accepted by 34th Conference on Neural Information Processing Systems (NeurIPS 2020) as spotlight.
2020/08: Wang Bingheng (PhD student), Nie Hongtuo (master student), and Sia Han Wei (master student) joined our group. Welcome!
2020/07: New manuscript on Q-learning acceleration: "Momentum Q-learning with Finite-Sample Convergence Guarantee"
2020/04: I joined National University of Singapore as an Assistant Professor in the CISR area (control, intelligent systems, and robotics). Thanks to my amazing colleagues at Aptiv/Motional for the fruitful collaborations over the past 2 years!