Research

Publications

• A General Framework for Learning Mean-Field Games.
  Xin Guo, Anran Hu, Renyuan Xu, and Junzi Zhang (alphabetical), 2020.
  Accepted, Mathematics of Operations Reseach.
  

• On the Global Convergence of Momentum-based Policy Gradient.
  Yuhao Ding, Junzi Zhang, and Javad Lavaei, 2021.
  To appear, International Conference on Artificial Intelligence and Statistics (AISTATS), 2022.
  

• Theoretical Guarantees of Fictitious Discount Algorithms for Episodic Reinforcement Learning and Global Convergence of Policy Gradient Methods.
  Xin Guo, Anran Hu and Junzi Zhang (alphabetical), 2021.
  To appear, AAAI Conference on Artificial Intelligence (AAAI), 2022.

• A Markov Regime Switching Model for Ultra-Short-Term Wind Power Prediction based on Toeplitz Inverse Covariance Clustering.
  Hang Fan, Xuemin Zhang, Shengwei Mei, Junzi Zhang, 2021.
  Frontiers in Energy Research (2021).
  

• Sample Efficient Reinforcement Learning with REINFORCE.
  Junzi Zhang, Jongho Kim, Brendan O'Donoghue, and Stephen Boyd, 2020.
  AAAI Conference on Artificial Intelligence (AAAI), 2021.
  

• Anderson Accelerated Douglas-Rachford Splitting.
  Anqi Fu*, Junzi Zhang*, and Stephen Boyd, 2019.
  SIAM Journal on Scientific Computing, 42.6 (2020): A3560–A3583.
  

  • a2dr: open-source Python solver for prox-affine distributed convex optimization.
    

• Learning Mean-Field Games. [Published version] [Slides]
  Xin Guo, Anran Hu, Renyuan Xu, and Junzi Zhang (alphabetical), 2019.
  Neural Information Processing Systems (NeurIPS), 2019.
  

• Globally Convergent Type-I Anderson Acceleration for Non-Smooth Fixed-Point Iterations. [Code]
  Junzi Zhang, Brendan O'Donoghue, and Stephen Boyd, 2018.
  SIAM Journal on Optimization, 30.4 (2020): 3170–3197.
  

  • (Partial) implementation in SCS 2.x & standalone package in C (with Python interface).
    

• Robust Super-Level Set Estimation using Gaussian Processes. [Published version] [Slides]
  Andrea Zanette*, Junzi Zhang*, and Mykel J. Kochenderfer.
  European Conference on Machine Learning (ECML), Oral, 2018.

Preprints & Working Papers

• Beyond Exact Gradients: Convergence of Stochastic Soft-Max Policy Gradient Methods with Entropy Regularization.
  Yuhao Ding, Junzi Zhang, and Javad Lavaei, 2021. (Submitted)

• A Continuous-Time Viewpoint of Broyden’s Methods.
  Honglin Yuan*, Junzi Zhang*, and Stephen Boyd, 2020. (Working paper)
  

  • Invited Talks: INFORMS Annual Meeting (Oct. 22) & Operations Research Seminar @ BICMR (Dec. 30), 2019.
    

• Consistency and Computation of Regularized MLEs for Multivariate Hawkes Processes.
  Xin Guo, Anran Hu, Renyuan Xu, and Junzi Zhang (alphabetical), 2018. (Submitted)
  Short version appeared in NeurIPS 2018 Workshop on Causal Learning.
  

Miscellaneous Writings

• Information-Directed Sampling for Reinforcement Learning. [Poster]
  Junyang Qian* and Junzi Zhang*, 2017.
  MS&E 338 course project supervised by Prof. Benjamin Van Roy & Dr. Abbas Kazerouni.

• Particle Filter Network: A Model-free Approach for POMDP. [Slides]
  Pengfei Gao* and Junzi Zhang*, 2018.
  AA 229 course project supervised by Prof. Mykel J. Kochenderfer.
  

• Ruminating Neural Networks for Sequence Modeling. [Link]
  Hao Sheng, Jin Xie, and Junzi Zhang (alphabetical), 2017.
  CS 224N Final Project Prize Winners.

Ph.D. Thesis

• Stabilizing Anderson Mixing for Accelerated Optimization.
  Junzi Zhang, 2021.
  [Note] Section 2.3 contains a correction and generalization of the restarting strategy in the original paper.