Σεμινάριο Τμήματος & CEID social hour
Oμιλητής: Konstantinos Chatzilygeroudis, Adjunct Faculty, CEID, U. Patras
Τίτλος: Micro-Data Reinforcement Learning for Adaptive Robots
Ημερομηνία-χώρος: Παρασκευή 26 Μαρτίου, 3-5μμ
Zoom link: https://upatras-gr.zoom.us/j/92569423452?pwd=SnZ3T0FmNEs0TCtiNjZXNjl6UVBuZz09
Περίληψη: Robots have to face the real world, in which trying something might take seconds, hours, or even days. Unfortunately, the current state-of-the-art learning algorithms (e.g., deep learning) rely on the availability of very large data sets. In this talk, we explore approaches that tackle the challenge of learning by trial and error in a few minutes on physical robots (we call this challenge “micro-data reinforcement learning”). We will first see how we can use simulators to generate behavior repertoires in order to develop algorithms that allow complex robots to quickly recover from unknown circumstances (e.g., damages or different terrain) while completing their tasks and taking the environment into account. In particular, we will see how a physical damaged hexapod robot can recover most of its locomotion abilities in an environment with obstacles, and without any human intervention, using this type of algorithms. Next, we will discuss how model-based reinforcement learning (RL) algorithms can be adapted so that we can use them on real physical robots. In particular, we will discuss (1) methods that leverage multi-core CPUs to enable fast computational times, and (2) how we can “scale” model-based RL methods to high-dimensional robots. More concretely, we will showcase algorithms that are able to find high-performing walking policies for a physical damaged hexapod robot (48D state and 18D action space) in less than 1 minute of interaction time. Next, we will present methods that aim to incorporate learning methods inside traditional control architectures. In particular, we will present work towards (a) incorporating data-driven methods into QP-based controllers, and (b) combining RL with stable dynamical system-based policies for efficient, stable and safe RL on physical systems. Finally, we will discuss current work towards efficient exploration for RL in robotics.
Σχετικά με τον ομιλητή: Dr. Konstantinos Chatzilygeroudis received the B.Sc. and M.Sc. degree in computer science and engineering from the University of Patras, Patras, Greece, in 2014, and the Ph.D. degree in robotics and machine learning from Inria Nancy-Grand Est, France and the University of Lorraine, Nancy, France in 2018. From 2018 to 2020 he was a Postdoctoral Fellow with the LASA Team with the Swiss Federal Institute of Technology Lausanne (EPFL), Lausanne, Switzerland. He is currently Leader of the R&D Computer Vision Team at Metargus, a pre-seed funded start up (based in Patras, Greece), building a cutting-edge basketball coaching tool to provide coaches with insights far beyond traditional analytics. He is also teaching under-graduate courses related to Artificial Intelligence and Robotics at the Computer Engineering & Informatics Department of University of Patras. He is currently serving as an Associate Editor at the International Conference on Intelligent Robotics (IROS), and participating in the organization committee (as a Virtual Chair) of the International Conference on Robot Learning (CoRL). His research interests include the area of artificial intelligence and focus on reinforcement learning and fast robot adaptation.