Random synaptic feedback weights support error backpropagation for deep learnings

Presenter: Sejnowski-Hinton Award 2025

Exhibit Hall F, 2pm

What is “Neruo AI/neuroai”?

Compute gradient using chain rule to perform the backprop