Intelligent Systems

Autonomous Learning Group

Max Planck Research Group for Autonomous Learning

We are interested in autonomous learning, that is how an embodied agent can determine what to learn, how to learn, and how to judge the learning success. In particular, we focus on learning to control a robotic body in a developmental fashion. Artificial intrinsic motivations are a central component that we develop using information theory and dynamical systems theory. We work on reinforcement learning, representation learning, and internal model learning.

Video

We are part of the MPI IS video series of short films presenting scientists' research projects in an understandable way.

In this clip we talk about the mission of our group and show two projects that we are working on:

Georg Martius, group leader of the Autonomous Learning Group, elaborates on the idea of developing robots as playful, autonomously learning machines. The vision is that robots learn the structure of their environments, rather than pre defined supervised tasks.

Cristina Pinneri, PhD student and fellow of the Center for Learning Systems (CLS), presents her project on autonomous, independent exploration of correlated motion patterns. The key idea is to provide robots with a fundamental self-understanding of its sensorimotor system and let it explore the environment in a self-organized way.

Sebastian Blaes, PhD student and fellow of the International Max-Planck Research School for Intelligent Systems (IMPRS-IS), talks about how robotic interactions with the environment can be learned. The focus of his project is on learning hierarchical task structures.