General Information

Full Name Stephen Chung
Languages English, Cantonese, Mandarin


  • 2022 - Now
    PhD in Engineering
    University of Cambridge, Cambridge, UK
  • 2019 - 2021
    MS in Computer Science
    University of Massachusetts Amherst, US
  • 2012 - 2016
    BS in Quantitative Finance
    University of Hong Kong, Hong Kong

Research Experience

  • 2022 - Now
    Learning to Plan and Act (Supervisor - David Krueger)
    University of Cambridge, Cambridge, UK
    • Investigated methods to train reinforcement learning agents to learn how to plan.
    • Studied methods to interpret the plans of reinforcement learning agents for detecting misaligned behaviour.
  • 2020 - 2022
    Training Neural Networks by Reinforcement Learning (Supervisor - Andrew Barto)
    University of Massachusetts Amherst, US
    • Investigated methods to train a deep artificial neural network without backpropagation based on reinforcement learning.
    • Proposed algorithms that can train a deep neural network to solve reinforcement learning tasks efficiently without any backpropagation.
  • 2020 - 2022
    Turing Completeness of Recurrent Neural Networks (Supervisor - Hava Siegelmann)
    University of Massachusetts Amherst, US
    • Proposed new theorems and constructed proof on the Turing completeness of bounded-precision recurrent neural networks.

Academic Interests

  • Reinforcement learning
    • Model-based RL, multi-agent RL, and hierarchical RL
  • Biologically-inspired machine learning
    • Spiking neural networks and biologically-inspired learning rules
  • AI alignment
    • Interpretability and detecting misaligned behaviour

Other Interests

  • Hobbies: Tennis, Reading, Poker, Soul-like and survival games