AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
2min chapter
061: Interpolation, Extrapolation and Linearisation (Prof. Yann LeCun, Dr. Randall Balestriero)
Machine Learning Street Talk (MLST)
Neural Networks Can't Extrapolate Outside of the Training Range
The only thing you need someo is to have the orientation of the axis where you will extrapolate, to be somewhat a line in the autagonal space of that yparplane. Once you have that, then you will generalize even thogh. But i still want to linger on this point of why you said or hears, just a random projection of the data. Here's a trained resinat projection. What i took from that is irrespective. It's all in an extraplative regime. I don't think people sull expect bad performance because you are not in a interpolation regime.
00:00
Transcript
Episode notes
Get the Snipd
podcast app
Unlock the knowledge in podcasts with the podcast player of the future.
AI-powered
Discover
highlights
Listen to the best highlights from the podcasts you love and dive into the full episode
Save any
moment
Hear something you like? Tap your headphones to save it with AI-generated key takeaways
Share
& Export
Send highlights to Twitter, WhatsApp or export them to Notion, Readwise & more
AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
Discover
highlights
Listen to the best highlights from the podcasts you love and dive into the full episode