📞 +49 30 314 29873
Mathias received his BSc in Bioinformatics from the Free University in Berlin and his MSc in Computational Neuroscience from the Technical University Berlin. During his studies, he was a visiting researcher with the Adaptive Behaviour and Coginition group at Plymouth University, UK, and the Adaptive Systems lab at Humboldt Universität zu Berlin which sparked his interest in representation learning and embodied cognition. He joined the Modelling of Cognitive Processes group for his PhD in February 2017 and is now working to understand the interplay between action and perception in reinforcement learning agents.
Sensory Representations and Reinforcement Learning
There is a long-standing interest in understanding the sensory processing of biological and artifical networks of neurons in terms of representations - computations of the raw sensory variables such as light pulses or pixel intensities into abstract variables that can be identified with meaningful variables of the environment, such as spatial position of an agent. Despite impressive recent progress in learning representations with Deep Neural Networks, it is as yet not well understood what characterizes good representations. We understand good representations as representations that generalize well across tasks and environments and approach the problem of learning them with a combination of spectral techniques such as Slow Feature Analysis and Deep Reinforcement Learning.