Jikun Kang

Contact me at firstname dot lastname at mail dot mcgill dot ca


Room 305

McConnel Building

McGill University

I am a Ph.D. candidate at McGill University. I am fortunate to be advised by Prof. Xue (Steve) Liu.

My research interest lies in the domain of Reinforcement Learning, Meta-RL and Applied RL. Recently, I’m working on improving RL policy generalization and sample-efficiency.


Dec 23, 2023 Our latest paper LaFFi: Leveraging Hybrid Natural Language Feedback for Fine-tuning Language Models has been accepted to AAAI 2024 Human-Centric Representation Learning.
May 28, 2023 Thrilled to see that my recent research with MSR has garnered attention from the community.twitter
Jan 18, 2023 Happy to share our paper “Multi-agent Attention Actor-Critic Algorithm for Load Balancing in Cellular Networks” was accepted at IEEE International Conference on Communications (ICC) 2023!
Sep 10, 2022 I’m thrilled to announce that our latest paper “Learning Multi-Objective Curricula for Robotic Policy Learning” has been accepted to CoRL 2022.
Aug 2, 2022 Our latest paper about how to improve policy generalization in real-world application (5G networking) has been accepted to GlobeCOM2022!

selected publications

  1. CoRL
    Learning Multi-Objective Curricula for Robotics Policy Learning
    Kang, Jikun, Liu, Miao, Gupta, Abhinav, Pal, Chris, Liu, Xue, and Fu, Jie
    CoRL 2022
  2. Globecom
    A Generalized Load Balancing Policy With Multi-Teacher Reinforcement Learning
    Kang, Jikun, Wang, Ju, Hu, Chengming, Liu, Xue, and Dudek, Gregory
    In GLOBECOM 2022
  3. ICC
    Hierarchical Policy Learning for Hybrid Communication Load Balancing
    Kang, Jikun, Chen, Xi, Wu, Di, Xu, Yi Tian, Liu, Xue, Dudek, Gregory, Lee, Taeseop, and Park, Intaik
    In ICC 2021
  4. Globecom
    Load Balancing for Communication Networks via Data-Efficient Deep Reinforcement Learning
    Wu, Di, Kang, Jikun, Xu, Yi Tian, Li, Hang, Li, Jimmy, Chen, Xi, Rivkin, Dmitriy, Jenkin, Michael, Lee, Taeseop, Park, Intaik, Liu, Xue, and Dudek, Gregory
    In GLOBECOM 2021
  5. NeurIPS
    Multi-Agent Option Critic Architecture
    Goupta, Abhinav, Chakravorty, Jhelum, Kang, Jikun, Liu, Xue, Pal, Christopher, and Precup, Doina
    In Neural Information Processing Systems (NeurIPS) workshops in Deep Reinforcement Learning 2020
  6. ICC
    Multi-agent Attention Actor-Critic Algorithm for Load Balancing in Cellular Networks
    Kang, Jikun, Wu, Di, Wang, Ju, Liu, Xue, and Dudek, Gregory
    In ICC 2023
  7. preprint
    Think Before You Act: Decision Transformers with Internal Working Memory
    Kang, Jikun, Laroche, Romain, Yuan, Xindi, Trischler, Adam, Liu, Xue, and Fu, Jie
    arXiv preprint arXiv:2305.16338 2023