The Robot Brains Podcast cover image

Simon Knowles on pushing AI computing to the limit by rethinking chips

The Robot Brains Podcast

00:00

Neural Networks - The Transformer

We bet that the data structures in the end would need to be grafts of great complexity. I think actually the neural network world, that we didn't anticipate being quite so pervasive, is now moving in that direction. They may be at their alex net moment. W to me, we've had two alex net moments so far. We've had the alex net moment itself, and then we've had the break out of transform. Almost the second one was interesting. It was the fact that you that that structure could harness unsupervised learning that was interesting thing. And under the hood, the transformer is quite different from the previous types of nera that work.

Transcript
Play full episode

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!
App store bannerPlay store banner
Get the app