Eduardo Sebastián
|
InfoResearch Associate ContactOffice: SN04 William Gates Building, 15 JJ Thompson Avenue, Cambridge, CB3 0FD, UK |
Biography
I am a Postdoctoral Researcher in the ProrokLab, at the University of Cambridge. I obtained a Ph.D. in Computer Science and Systems Engineering from the Universidad de Zaragoza in 2025, supervised by Prof. Eduardo Montijano and Prof. Carlos Sagüés. I visited Prof. Nikolay A. Atanasov at the University of California, San Diego, in 2022 and 2024. I am a Fulbright Scholar, an RSS Pioneer, a DAAD AInet Fellow and a Research Associate of Corpus Christi college.
Research Interests
I work on the algorithmic foundations of collective intelligence to build teams of robots that reason, collaborate and solve complex tasks with a minimal energy footprint.
News
Oct 2026 I'm co-organizing a very cool workshop: One Idea, Many Names? Internal Models of the World in Natural and Artificial Systems. It will take place at SAB'26 and is organized together with the RBO Lab at TU Berlin.
Aug 2026 I've received the SCIE-BBVA Foundation Research Award for young researchers, given to the top six computer scientists under 30 in Spain (first ever awarded to a roboticist!).
Jun 2026 One paper accepted at IEEE RAL. It studies how to command collaborative teams of robots through natural language in real-time.
Jun 2026 One paper accepted at IROS'26 on how to use self-attention mechanisms to solve model-free distributed nonlinear multi-agent games.
Jun 2026 New preprint. We study what we argue is the most fundamental factorization in robot learning: separating the world from the task, showing evidence for out-of-distribution generalization due to compositionality.
May 2026 New preprint. We propose a new perspective on behavioral diversity for MARL in which events play a key role, showing emergent cooperation patterns and improved performance and generalization.
Apr 2026 I have been selected as RSS Pioneer 2026!
Mar 2026 New preprint. We study wake effects suffered by robots interacting in dynamic regimes in a fluid medium, concluding that memory and transport delay prediction are key for accurate modelling.
Feb 2026 My PhD thesis has been selected as finalist for the 2025 IEEE RAS Technical Committee on Multi-Robot Systems Best Dissertation Award.
Dec 2025 New preprint. We introduce a neuromorphic control framework that leverages cognitive models of decision-making to resolve spatial indecision, enabling robots to break indecision directly from visual inputs.
Sep 2025 Two new preprints. The first one studies how to command collaborative teams of robots through natural language in real-time, and the second one proposes a novel policy gradient method with self-attention for model-free distributed nonlinear multi-agent games.
Sep 2025 Our paper Estimation of Semiconductor Power Losses Through Automatic Thermal Modeling has been awarded a 2024 IEEE Transactions on Power Electronics (TPEL) Second Place Prize Paper Award.
Sep 2025 One paper and one poster accepted. The paper studies curriculum learning techniques for learning scalable, distributed multi-robot policies and will be presented at ROBOT'25. The poster introduces a novel neuromorphic control for autonomous vision-guided flight and will be presented at NERC'25.