Timezone: »

Deep Learning without Weight Transport
Mohamed Akrout · Collin Wilson · Peter Humphreys · Timothy Lillicrap · Douglas Tweed

Tue Dec 10 10:45 AM -- 12:45 PM (PST) @ East Exhibition Hall B + C #102

Current algorithms for deep learning probably cannot run in the brain because they rely on weight transport, where forward-path neurons transmit their synaptic weights to a feedback path, in a way that is likely impossible biologically. An algorithm called feedback alignment achieves deep learning without weight transport by using random feedback weights, but it performs poorly on hard visual-recognition tasks. Here we describe two mechanisms — a neural circuit called a weight mirror and a modification of an algorithm proposed by Kolen and Pollack in 1994 — both of which let the feedback path learn appropriate synaptic weights quickly and accurately even in large networks, without weight transport or complex wiring. Tested on the ImageNet visual-recognition task, these mechanisms outperform both feedback alignment and the newer sign-symmetry method, and nearly match backprop, the standard algorithm of deep learning, which uses weight transport.

Author Information

Mohamed Akrout (University of Toronto)
Collin Wilson (University of Toronto)
Peter Humphreys (Deepmind)
Timothy Lillicrap (DeepMind & UCL)
Douglas Tweed (University of Toronto)

More from the Same Authors