DCN seminar by Prof.dr. Sander M. Bohté, CWI ML group

About the speaker

Prof. Bohté develops bio-inspired an biologically plausible neural networks. The challenge is to develop insights from neuroscience into usefully computing neural networks, and to bring machine learning insights into models of how neurons in the brain compute. I particularly focus on continuous-time neural information processing, where time is an explicit dimension of the problem domain. This includes networks of spiking neurons, models of attention, predictive coding, interactive neural cognition, supervised neural learning, and deep reinforcement learning methods.

• A key research interest is scaling spiking neural networks to high accuracy and large-scale networks. Recent work includes this paper in Nature Machine Intelligence and this this pre-print on large-scale efficient online learning. In the NWA ACT project Bohté applies these insights to human-machine interaction for autonomous vehicles.
• In collaboration with Pieter Roelfsema at NIN, Bohté works on biologically plausible deep learning based on the assumption that animals learn behaviour through reinforcement learning. See "Brainprop".
• In more applied machine learning efforts, Bohté works on the application of AI in computational physics (with Nikolaj Mucke, Benjamin Sanderse and Kees Oosterlee).
• Bohté holds an appointment as a part-time professor of Computational Neuroscience at the University of Amsterdam and as a part-time professor of Bio-Inspired Deep Learning at the University of Groningen.