Brain-Aligned Generative Models

Generative AI has advanced rapidly, yet generative processes in the brain remain a blind spot for neuroscience. Adrien Doerig from Freie Universität Berlin is building brain-aligned generative models that bring neuroscience and machine learning together – a path toward AI that is both more interpretable and more efficient.

Biography

Can a language model predict what your brain is doing? Adrien Doerig is a professor of cognitive computational neuroscience at Freie Universität Berlin and the Bernstein Center for Computational Neuroscience, where his lab bridges neuroscience and machine learning. A lead author of the influential “neuroconnectionist” research programme, he recently showed that large language models can predict how the human brain responds to visual scenes — and is now building brain-aligned generative models.