State of the Art Neural Networks - Neural architecture search (NAS)

[@Diallo268]

Good video. My only issue is that the volume is really low. You could normalize it to something like 95% or above.

[@titan_x_alpha5196]

I study in LPU University and I got an assignment to watch full video and then explain the video in 250 words, soo if anyone who watched can u please explain me in 250 words 😅

[@JustSomeElectricalEngineer]

Isn't NAS a subfield of HPO? I thought NAS was only concerned with architectural choices, whereas there are plenty of HParams that do not define the architecture.

[@sudhakarm4573]

well understood.. good clarity

[@olivierramier8335]

Couldn't the architectures being tested be AI generated as well, using as input a description of the task and the constraints? It could be trained on what data around what architecture works best for which task. The intuition of experts when designing an architecture comes from somewhere...

[@DG-rv9ph]

amazing work

[@siddhawan5190]

This is great 👍

[@aminerachid4393]

The speaker is nervous and it shows

[@Jack-lg9mq]

Most people get nervous presenting. Pointing it out is unproductive.