Michael Matthews

Michael Matthews

PhD Student

University of Oxford

Biography

I am a first-year PhD student at the University of Oxford studying reinforcement learning in open-ended settings. I am supervised by Dr. Jakob Foerster as part of FLAIR.

I completed my MSc in Machine Learning at University College London, where I worked on transfer learning in RL at UCL DARK Lab under the supervision of Prof. Tim Rocktäschel and Prof. Edward Grefenstette. Before this, I obtained my BA in Computer Science from the University of Cambridge.

Before starting my PhD I worked for 2 years as a Machine Learning Researcher at VivaCity, where I worked on using RL for traffic control. During my undergrad I completed software engineering internships at G-Research and PlayFusion, where I worked on internal trading software and RL for video games.

Interests
  • Open-Endedness
  • Reinforcement Learning
  • Exploration
Education
  • DPhil Engineering Science, 2027

    University of Oxford

  • MSc Machine Learning, 2021

    University College London

  • BA Computer Science, 2020

    University of Cambridge

Publications

(2024). Craftax: A Lightning-Fast Benchmark for Open-Ended Reinforcement Learning. ICML 2024 (Spotlight).

PDF Cite

(2024). Policy-Guided Diffusion. RLC 2024.

PDF Cite

(2024). Refining Minimax Regret for Unsupervised Environment Design. ICML 2024.

PDF Cite

(2022). Hierarchical Kickstarting for Skill Transfer in Reinforcement Learning. CoLLAs 2022.

PDF Cite