Scholar
Timon Willi
Google Scholar ID: Dn-udzAAAAAJ
University of Oxford
Citations & Impact
Citations
439
H-index
9
i10-index
9
Publications
17
Co-authors
21
list available
Publications
Resume (English only)
Academic Achievements
- ICML 2024: "Mixtures of Experts Unlock Parameter Scaling for Deep RL" (Spotlight, top 3.5%)
- AAMAS 2024: Multiple papers including "Scaling Opponent Shaping to High Dimensional Games" and "Analysing the Sample Complexity of Opponent Shaping" (both oral presentations)
- ICML 2022: Published "Model-Free Opponent Shaping" and "COLA: Consistent Learning with Opponent-Learning Awareness"
- NeurIPS 2019 BDL Workshop: "Recurrent Neural Processes" (oral presentation)
- Developed open-source projects JaxMARL (multi-agent RL library in JAX) and Pax (scalable opponent shaping framework)
- Co-developed Jafar: an open-source JAX reimplementation of Genie
