On the Connection between Neural Networks and Kernels: a Modern Perspective -Simon Du

Workshop on Theory of Deep Learning: Where next?
Topic: On the Connection between Neural Networks and Kernels: a Modern Perspective
Speaker: Simon Du
Affiliation: Member, School of Mathematics
Date: October 16, 2019
For more video please visit video.ias.edu

[@luke2642]

It feels intuitive that global average pooling should have some value... like to Hinton's capsule networks, except the capsules are blasted to pieces and diluted into an ensemble, rather than discrete.

[@lukistrela]

at 4:55 there is a promise of complexity measure that explains what VC-dim cannot. At the end of the talk, I fail to see what that measure is.

[@lukistrela]

at 10:40, under Trajectory of predictions, the sum of angle brackets -- there are two scalars inside scalar brackets???

[@bingeltube]

Why are the presentation slides not provided? This talk is very math heavy

[@piergiorgiobenenati3698]

Amazing work, but sorry... I cannot understand the speaker's English most of the time. The very frequent mismatch between the UA-cam automatic subtitles and the speaker's words corroborates my opinion.