Making robots learn in the real-world
Most of nowadays research is focused on learning in simulation. Learning in simulation, however, is not scalable, as it requires the design of new simulator environments for each task. Ultimately, we want to have robots that learn in the real world, similarly to animals and humans. However, learning in the real world is challenging. My research focuses on providing helpful representation and abstraction for embodied agents. I also focus on theoretical aspects of reinforcement learning to provide higher efficiency.
For more information about my research interest visit samueletosatto.com
I currently teach
See Google Scholar.
2021–2022 Postdoctoral Researcher, University of Alberta, Canada
2017–2020 Ph.D., Computer Science, Technical University of Darmstadt
2013–2017 M.Sc., Software Engineering II, Polytechnic University of Milan.
2009–2012 B.Sc., Software Engineering I, Polytechnic University of Milan.