Javier García is a faculty member at the University Carlos III of Madrid, affiliated with the Department of Computer Science within the School of Engineering. His research lies at the intersection of artificial intelligence, reinforcement learning, and robotics, with a strong emphasis on safety, transferability, and real-world applications. His research interests include: Safe and robust reinforcement learning Transfer and multi-task learning in AI Adversarial attacks and defenses in deep RL Automated planning and decision-making under uncertainty Applications in social assistive robotics and financial systems The recent trend in his publications shows a focus on formalizing similarity between decision-making tasks, improving robustness against corrupted rewards and adversarial policies, and applying AI techniques to domains such as geriatric care and automated market making. His work often involves collaboration with Fernando Fernández and other researchers in Spain. He has contributed significantly to frameworks like CLARC for comprehensive geriatric assessment using robots, and has developed tools such as TIMIPLAN for transportation tasks. His research bridges theoretical AI with practical deployment in cognitive robotics and business simulators. Notable scientific contributions include a comprehensive survey on safe reinforcement learning (JMLR 2015), foundational work on policy reuse, and the development of taxonomies for MDP similarity. He has secured research funding through collaborative projects, though specific grants are not detailed in the provided text. He has advised several students and junior researchers, evident from co-authored publications, though specific names are not listed. He is part of research teams working on lifelong learning technologies and cognitive robot systems.
