profile_picture
HENRIQUE DONÂNCIO
Postdoc, INRIA Grenoble

I am in the job market! Meanwhile, I keep learning and posting about it on the Blog. Check it out!

I am a former Postdoc Researcher at the Statify team of Inria Grenoble, where I worked with Florence Forbes. My research is related to Deep Reinforcement Learning. I am mostly interested in developing algorithms and techniques for agents to learn from a continuous and non-stationary data stream.

Before I did my Ph.D. at LITIS lab as part of the IoT.H2O project under Prof. Laurent Vercouter’s supervision. My research was focused on enhancing the learning process by using logged data or structuring it through curriculum learning.

The research topics I am interested in include exploiting knowledge from static datasets (learning from demonstrations), transfer learning, distributional RL, multi-armed bandits, meta-learning, and applied RL. Lately, I have been interested in continual learning, multi-agent RL and optimisation.

Interests

  • (Deep) Reinforcement Learning
  • Continual Learning
  • Machine Learning
  • Optimisation
  • Multi-agent Systems

Academia

INSA Rouen Normandie
2019 - 2023
Ph.D. Artificial Intelligence
supervised by Prof. Laurent Vercouter
University of São Paulo (USP)
2017 - 2019
M.Sc. Computer Science
supervised by Prof. Anarosa A. F. Brandão
Federal University of Rio Grande (FURG)
2011 - 2017
B.Sc. Computer Engineering
supervised by Prof. Graçaliz Dimuro and Prof. Diana Adamatti

News

You can find here my latest news

Papers and Preprints

An analysis of distributional reinforcement learning with Gaussian mixtures, 2025, TMLR (accepted 2026)
Mathis Antonetti , Henrique Donâncio , Florence Forbes
Dynamic Learning Rate for Deep Reinforcement Learning: A Bandit Approach , 2024, arXiv preprint (under review)
Henrique Donâncio , Antoine Barrier , Leah F. South , Florence Forbes
The Pump Scheduling Problem: A Real-World Scenario for Reinforcement Learning, 2022, arXiv preprint
Henrique Donâncio , Laurent Vercouter , Harald Roclawski
Safety through Intrinsically Motivated Imitation Learning, 2022, ALA@AAMAS 2022
Henrique Donâncio , Laurent Vercouter

Projects

See my gitlab for the complete list
The Pump Scheduling Problem