Yuhao Ding 丁宇昊
 |
I got my Ph.D. degree in Industrial Engineering and Operations Research from University of California, Berkeley, advised by Javad Lavaei. My research has been focused on the interdisciplinary problems in control, reinforcement learning, optimization and statistical learning. Most recently, I work on the reinforcement learning under the safety requirements and the time-varying environments. Email: yuhao_ding [@] berkeley [DOT] edu, |
News
September 2023: Two papers to appear in Conference on Neural Information Processing Systems (NeurIPS): Scalable Primal-Dual Actor-Critic Method for Safe Multi-Agent RL with General Utilities and Tempo Adaption in Non-stationary Reinforcement Learning.
August 2023: New paper on optimization landscape of dynamic programming: The Landscape of the Optimal Control Problem: One-shot Optimization Versus Dynamic Programming.
July 2023: New paper to appear in MILETS workshop at Knowledge Discovery and Data Mining Conference (KDD): Dynamic Ensembling for Probabilistic Time Series Forecasting via Deep Reinforcement Learning.
May 2023: New paper on multi-agent reinforcement learning: Scalable Primal-Dual Actor-Critic Method for Safe Multi-Agent RL with General Utilities.
April 2023: I defended my PhD.
March 2023: The paper Learning-to-Learn to Guide Random Search: Derivative-Free Meta Blackbox Optimization on Manifold to appear in Learning for Dynamics & Control Conference (L4DC) as an oral presentation.
March 2023: The paper Local Analysis of Entropy-Regularized Stochastic Soft-Max Policy Gradient Methods to appear in European Control Conference (ECC).
February 2023: I gave a talk on safe reinforcement learning online seminar.
January 2023: Our paper A CMDP-within-online framework for Meta-Safe Reinforcement Learning is accepted as a spotlight in International Conference on Learning Representations (ICLR).
January 2023: Our paper Scalable Multi-Agent Reinforcement Learning with General Utilities is accepted for American Control Conference (ACC).
January 2023: Our paper Non-stationary Risk-sensitive Reinforcement Learning is accepted as an oral in AAAI Conference on Artificial Intelligence (AAAI).
November 2022: New Paper on meta learning: Learning-to-Learn to Guide Random Search: Derivative-Free Meta Blackbox Optimization on Manifold.
November 2022: Three papers to appear in AAAI Conference on Artificial Intelligence (AAAI): Non-stationary Risk-sensitive Reinforcement Learning, Convex Constrained Markov Decision Processes, and Non-stationary Constrained Markov Decision Processes.
October 2022: I organized a session on ‘‘Recent Advances in Provably Efficient Reinforcement Learning with Safety Guarantee’’ at the 2022 INFORMS Annual Meeting. We were excited and honored to have a great team of speakers (alphabetical): Dongsheng Ding, Ming Jin, Alec Koppel and Donghao Ying.
October 2022: New paper on reinforcement learning: Scalable Multi-Agent Reinforcement Learning with General Utilities.
May 2022: I started the internship at Amazon AWS AI Lab as an applied scientist intern.
May 2022: New paper on reinforcement learning: Provable Guarantees for Meta-Safe Reinforcement Learning.
May 2022: New paper on reinforcement learning: Policy-based Primal-Dual Methods for Convex Constrained Markov Decision Processes.
May 2022: New paper on reinforcement learning: Non-stationary Risk-sensitive Reinforcement Learning: Near-optimal Dynamic Regret, Adaptive Detection, and Separation Design.
May 2022: I received the Marshall-Olivier-Rosenberger Fellowship given by UC Berkeley.
Jan 2022: New paper on reinforcement learning: Provably Efficient Primal-Dual Reinforcement Learning for CMDPs with Non-stationary Objectives and Constraints.
Jan 2022: Two papers to appear in International Conference on Artificial Intelligence and Statistics (AISTATS): A Dual Approach to Constrained Markov Decision Processes with Entropy Regularization, and On the Global Optimum Convergence of Momentum-based Policy Gradient.
Oct 2021: I gave a guest lecture on model-free reinforcement learning for IEOR 268 Applied Dynamic Programming at UC Berkeley.
Oct 2021: I gave a talk for the session on Recent Advances in Data Efficient Reinforcement Learning with Policy Gradient Methods at the 2021 INFORMS Annual Meeting.
Oct 2021: Two new papers posted on arXiv. In these papers, we derive the first set of global convergence results for stochastic policy gradient methods with momentum and entropy.
Oct 2021: New paper on constrained Markov decision processes: A Dual Approach to Constrained Markov Decision Processes with Entropy Regularization.
July 2021: The paper Structured Projection-free Online Convex Optimization with Multi-point Bandit Feedback to appear in 60th Conference on Decision and Control.
June 2021: The paper Time-variation in online nonconvex optimization enables escaping from spurious local minima has been conditionally accepted for IEEE Transactions on Automatic Control.
May 2021: I started the research internship at Microsoft research.
April 2021: The paper On the absence of spurious local trajectories in time-varying nonconvex optimization has been conditionally accepted for IEEE Transactions on Automatic Control.
January 2021: The paper Escaping spurious local minimum trajectories in online time-varying nonconvex optimization has been selected as a best student paper finalist for 2021 American Control Conference.