Jakob H. Macke is Professor for `Machine Learning in Science’ at the University of Tübingen, and Director of the Bernstein Center for Computational Neuroscience Tübingen. His group builds computational tools for analyzing neural and behavioural data, with the goal of understanding the principles of neural computation in biological and artificial neural systems.
I want to understand how populations of neurons collectively process sensory information and guide complex behaviour. To this end, we develop statistical methods and machine learning algorithms for analysing measurements of neural activity animal behaviour.
Advances in experimental techniques make it possible to measure structure and function of neural circuits at unprecedented scale and resolution. However, interpreting the complex data generated by these approaches has proven to be a difficult challenge which requires powerful analysis tools. At the same time, the field of machine learning is being revolutionised by ‘deep learning’ approaches, which use artificial neural networks for extracting structure from data. We are therefore faced with enormous opportunities for synergy between machine learning and neuroscience: To realize this potential, we pursue an inter-disciplinary approach combining elements from machine learning, computational neuroscience and engineering: We develop statistical models for analyzing high-dimensional data produced by modern imaging- and recording methods in neuroscience.
Closely collaborating with experimental laboratories, we aim to derive models that can explain neural measurements and adaptive animal behavior. Through a better understanding of the algorithms, abstractions and representations used by the brain, we want to contribute to the development of more powerful and flexible algorithms for artificial learning systems.