About me

I am currently a Ph.D. student advised by Porf. Tianyi Chen at RPI. I was fortunate to join Prof. Tianyi Chen’s group as the first Ph.D. student and grow with the research group till today. My research generally focuses on optimization algorihthms, bilevel learning and reinforcement learning.

I’m actively looking for industrial research internships. Email me for an updated CV.

Research

My research spans the areas of optimization algorithms and machine learning, covering the following topics:

  • Bilevel learning algorithms Bilevel learning is a general learning framework covering a wide range of topics–adversarial learning, hyper-parameter optimization, meta learning, etc. Our research focuses on the the theory foundation and algorithm design for bilevel training. We aim to advance and improve over the current single-level training paradigm with bilevel training framework.

  • Reinforcement learning Solving reinforcement learning problems with both online or offline policy optimization algorithms, along with recent focus on AI alginment with reinforcement learning from human feedback.

I have developed theories and implemented new algorithms in the area of bilevel/multi-objective learning and reinforcement learning over the years, thus gaining a mixure of theoretical and empirical skills. At the present stage, I am exploring enhancing language model alignment with reinforcement learning.

News

Publications

  • Joint Unsupervised and Supervised Training for Automatic Speech Recognition via Bilevel Optimization
    A.F.M. Saif, Xiaodong Cui, Han Shen, Songtao Lu, Brian Kingsbury, Tianyi Chen
    to appear in ICASSP 2024.

  • A Method For Bilevel Optimization With Convex Lower-level Problem
    Han Shen, Santiago Paternain, Gaowen Liu, Ramana Kompella, Tianyi Chen
    to appear in ICASSP 2024.

  • On Penalty-based Bilevel Gradient Descent Method
    Han Shen, Quan Xiao, Tianyi Chen
    ICML 2023. [arxiv] [code]

  • Towards Understanding Asynchronous Advantage Actor-critic: Convergence and Linear Speedup
    Han Shen, Kaiqing Zhang, Mingyi Hong, Tianyi Chen
    IEEE Transactions on Signal Processing. [arxiv]

  • Alternating projected SGD for equality-constrained bilevel optimization
    Quan Xiao, Han Shen, Wotao Yin, Tianyi Chen
    AISTATS 2023. [arxiv]

  • A Single-timescale Analysis for Stochastic Approximation with Multiple Coupled Sequences
    Han Shen, Tianyi Chen
    NeurIPS 2022 (oral). [arxiv]

  • Distributed Offline Policy Optimization Over Batch Data
    Han Shen, Songtao Lu, Xiaodong Cui, Tianyi Chen
    AISTATS 2022. [html]

  • Mitigating Gradient Bias in Multi-objective Learning: A Provably Convergent Approach
    Heshan D. Fernando, Han Shen, Miao Liu, Subhajit Chaudhury, Keerthiram Murugesan, Tianyi Chen
    ICLR 2022. [arxiv]

  • Adaptive Temporal Difference Learning with Linear Function Approximation
    Tao Sun, Han Shen, Tianyi Chen, Dongsheng Li
    IEEE Transactions on Pattern Analysis and Machine Intelligence. [arxiv]

  • Byzantine-resilient Decentralized Policy Evaluation with Linear Function Approximation
    Zhaoxian Wu, Han Shen, Tianyi Chen, Qing Ling
    IEEE Transactions on Signal Processing. [arxiv]

Services

Reviewer for

  • Advances in Neural Information Processing Systems (NeurIPS) 2022 & 2023
  • International Conference on Machine Learning (ICML) 2023 & 2024
  • International Conference on Learning Representation (ICLR) 2022
  • International Conference on Artificial Intelligence and Statistic (AISTATS) 2023 (top reviewer)
  • IEEE Transactions on Signal Processing (TSP)

Internships

IBM Research. (US) 05.2021 - 08.2021