Robots Learning Through Interactions

The acquisition and improvement of new motor skills represent one of the most relevant challenges in robotics. Complexity arises from interactions with the environment and with humans. A human teacher is always involved in the learning process, either directly (by providing data) or indirectly (by defining the optimization criterion). This raises the question: how can interactions with the human teacher be best leveraged to make the learning process efficient and effective?
Professor Jens Kober will present several methods that, together with his team, he has developed in the fields of supervised learning, imitation learning, reinforcement learning, and interactive learning. These will be illustrated with real robot experiments, ranging from playful examples (such as ball-in-a-cup) to more practical applications (retail environments).

The Ellis Turin Talk seminar series resumes with the event Robots Learning Through Interactions, delivered by Professor Jens Kober (Delft University of Technology). The event is organized in collaboration with the Artificial Intelligence Hub of Politecnico di Torino and the Vandal research group of the Department of Control and Computer Engineering-DAUIN.

The meeting will be held on September 10, 2025, at 4:30 p.m., and can be attended in the following two ways:

• in person, at Sala Grande Covivio, Corso Ferrucci 112;
• online via Zoom, by clicking on this link.

For further information, please contact Alessandra Calosso.