10 October, 2019 @ 6:30 PM – Pi 6 (via Indonesia 23, Rome)
Can we use machine learning to fuel scientific discoveries? To answer this question, Andrea Banino, a machine learning researcher working on artificial general intelligence at DeepMind, will share with us the results of a research project showing that deep learning techniques could be used to investigate the brain in new intriguing ways.
Andrea will illustrate a work in which his team first trained a recurrent network to perform path integration, then led it to the emergence of representations resembling grid cells, found in the mammalian brain. Those representations provide an effective basis for an agent to locate goals in challenging, unfamiliar, and changeable environments, optimizing the primary objective of navigation through deep reinforcement learning.
Join us to discuss how AI can be used to better understand how our brain works and to discover how neuroscience can be used to design AI algorithms.
Andrea Banino is a machine learning researcher working on artificial general intelligence at DeepMind.
His research focuses primarily on deep learning, deep reinforcement learning and neuroscience and in particular on memory and spatial navigation.
This seminar is part of the IAML tech talks, a serie of lectures co-organized and sponsored by the Italian Association for Machine Learning and Pi Campus, to promote discussion on topics of wide significance and societal impact, and bringing together researchers, professionals, and policy makers to discuss the future of machine learning in Italy.