Wasserstein Belief Updater was accepted at ICLR!
Raphael Avalos

Raphael Avalos

PhD Candidate



Hello! I am a PhD student at the AI LAB of VUB (University of Brussels), where I am supervised by Professors Ann Nowe and Diederik Roijers. The central focus of my research is the challenge of partial observability in both single-agent and multi-agent reinforcement learning (RL). My approach to addressing this issue is centered on leveraging additional data during the learning phase and potentially incorporating communication at the execution stage. My current main interests are in model-based RL and exploration strategies I welcome academic discussions, collaborations, or professional inquiries related to these areas. Feel free to contact me.

Download my resumé .

  • Artificial Intelligence
  • Reinforcement Learning
  • Partially Observable
  • Multi-Agent RL (collaboration, communication)
  • Model Based RL
  • Exploration
  • MSc in Applied Mathematics for AI, 2019

    Ecole Normale Superieur de Cachan

  • MSE in Computer Science, 2019

    Telecom Paris


Program Chair
October 2023 – May 2024 Auckland, New Zeland
Local Chair
January 2023 – September 2023 Brussels, Belgium
Vrije Universiteit Brussel (VUB) - AI Lab - FWO
PhD Candidate
May 2020 – Present Brussels, Belgium
PhD supervised by Prof. Ann Nowe and Dr. Diederik Roijers.
Vrije Universiteit Brussel (VUB) - AI Lab
Research Intern
January 2020 – April 2020 Brussels, Belgium
Working on Multi-agent Reinforcement Learning under the supervision of Prof. Ann Nowe.
INRIA - SequeL Team (now Scool)
Research Intern
April 2019 – October 2019 Lille, France
Worked on Hierarchical Reinforcement Learning (HRL) applied to the Rubik’s Cube under the supervision Dr. Florian Strub and Prof. Philippe Preux.


PhD Fellowship fundamental research

Journal Publications

(2022). Local Advantage Networks for Multi-Agent Reinforcement Learning in Dec-POMDPs. TMLR.

PDF Cite

Conference Publications

(2024). The Wasserstein Believer: Learning Belief Updates for Partially Observable Environments through Reliable Latent Space Models. ICLR 2024.

PDF Cite

(2023). Laser Learning Environment: A new environment for coordination-critical multi-agent tasks. BNAIC 2023.


(2022). Local Advantage Networks for Cooperative Multi-Agent Reinforcement Learning. AAMAS 2022.

PDF Cite

Workshop Publications

(2022). Autocurricula and Emergent Sociality from a Gene Perspective. ALA 2022 (AAMAS Workshop).

PDF Cite Video

(2022). Multi-agent RMax for Multi-Agent Multi-Armed Bandits. ALA 2022 (AAMAS Workshop).

PDF Cite Video


(2017). Dynamic Size Message Scheduling for Multi-Agent Communication under Limited Bandwidth. Preprint.

PDF Cite