The fields of neuroscience and AI have undergone significant transformation over the last ten years, thanks in large part to a rapid expansion in technological capacities that are permitting ever more complex experiments and computations to be carried out. Brain-like computing is becoming common in artificial intelligence and machine learning. But can brains and machines think alike?
How will neuroscience and AI shape each other’s futures?
This is a crucial question as our understanding of the brain increases, and the AI in our devices and machines increasingly permeate society, business and the economy. To map the road ahead, this symposium gathered global leaders in both neuroscience and artificial intelligence. Speakers not only described their work, but also shared their thoughts on the best path going forward.
(Click on any title below to link to video of the lecture)
Conference Program
Friday evening (Chaired by Anne Churchland)
7 pm — Session 1 — Brains and machines
David Heeger, NYU – ORGaNICs: A theory of working memory in brains and machines
Eve Marder, Brandeis – Surprising robustness and reliability in neuronal circuits
Yoshua Bengio, Université de Montréal – Bridging the gap between deep learning and neuroscience
Saturday morning (Chaired by David Heeger)
8 am — Session 2 — Biological and artificial mechanisms
Blaise Agüera y Arcas, Google – Learning locally and globally
Mu-ming Poo, Chinese Academy of Sciences – Synaptic plasticity and brain-inspired machine learning
Terry Sejnowski, Salk Institute – The global brain
Coffee break
10.30 am — Session 3 — Action
Leslie Pack Kaelbling, MIT – Making robots behave
Daniel Wolpert, Cambridge – Probabilistic models of sensorimotor control
Matteo Carandini, UCL – Testing the textbook model of brain function
Lunch break
Saturday afternoon (Chaired by Matteo Carandini)
2 pm — Session 4 — Cognition
Stanislas Dehaene, Collège de France – What is consciousness, and could machines have it?
Matthew Botvinick, DeepMind – Meta-learning in brains and machines
Coffee break
4.30 pm — Session 5 — Navigating and Remembering
David Tank, Princeton – Characterizing neural dynamics during navigation and decision-making
Sunday morning (Chaired by Tony Movshon)
8 am — Session 6 — Vision 1
Yann LeCun, Facebook – What are the principles of learning in newborns?
Jim DiCarlo, MIT – Reverse engineering visual intelligence
Eero Simoncelli, NYU – Perceptual implications of hierarchical visual models
Coffee break
10.30 am — Session 7 — Vision 2
Adrienne Fairhall, University of Washington – Rules of adaptation across cortex
Nicole Rust, University of Pennsylvania – Adaptation as a canonical mechanism for memory
Shimon Ullman, Weizmann Institute of Science – Image understanding beyond object recognition
Lunch break
Sunday afternoon (Chaired by Anne Churchland)
2 pm — Session 8 — Learning 1
Zach Mainen, Fundação Champalimaud – Serotonin and the regulation of neural inference and learning
Coffee break
4.30 pm — Session 9 — Learning 2
Greg Corrado, Google – Practical intelligence and skeptical minds
Josh Tenenbaum, MIT – Building machines that learn and think like people