Scholar
Roger Creus Castanyer
Google Scholar ID: E3y_txsAAAAJ
Mila/University of Montreal
Citations & Impact
Citations
82
H-index
5
i10-index
4
Publications
17
Co-authors
0
Publications
Resume (English only)
Academic Achievements
- ARM-FM: Automated Reward Machines via Foundation Models for Compositional Reinforcement Learning (preprint, October 2025); Stable Gradients for Stable Learning at Scale in Deep Reinforcement Learning (NeurIPS 2025, Spotlight, top 3% submissions); AI Research Scholarship (February 2025); Academic Excellence Scholarship (December 2024); Surprise-Adaptive Intrinsic Motivation for Unsupervised Reinforcement Learning (RLC 2024); RLeXplore: Accelerating Research in Intrinsically-Motivated Reinforcement Learning (TMLR 2024); Improving Intrinsic Exploration by Creating Stationary Objectives (ICLR 2024); PixelEDL: Unsupervised Skill Discovery and Learning from Pixels (Embodied AI workshop @ CVPR 2021); Unsupervised Skill-Discovery and Skill-Learning in Minecraft (Unsupervised Reinforcement Learning workshop @ ICML 2021); PiCoEDL: Discovery and Learning of Minecraft Navigation Goals from Pixels and Coordinates (Embodied AI workshop @ CVPR 2021); Integration of Convolutional Neural Networks in Mobile Applications (Workshop on AI Engineering @ ICSE 2021); Which Design Decisions in AI-enabled Mobile Applications Contribute to Greener AI? (Empiricial Software Engineering Journal 2022); Enhancing sequence-to-sequence modelling for RDF triples to natural text (WebNLG workshop).
Research Experience
- Research Intern @ Ubisoft LaForge (Montreal, Canada); Teaching Assistant @ University of Montreal (Montreal, Canada); Junior Data Scientist @ HP Inc (Barcelona, Spain); Research Assistant @ UPC (Barcelona, Spain); Basketball Coach @ Sagrada Familia Claror (Barcelona, Spain).
Education
- PhD: Mila / UdeM (Fall 2024 onwards), Supervisors: Pablo Samuel Castro and Glen Berseth; MSc: Mila Québec & University of Montréal (Montreal, Canada); BSc: Universitat Politècnica de Catalunya (UPC) (Barcelona, Spain).
Background
- Research Interests: Deep Reinforcement Learning, Foundation Models (LLMs, VLMs), AI Agents. Focus: Building general, autonomous agents that integrate the structured learning and adaptability of RL with the broad priors and reasoning abilities of foundation models, improving exploration, credit assignment, and skill discovery.
Miscellany
- From Barcelona, Spain, currently in Montreal, Canada.
Co-authors