AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
3min chapter
#85 Dr. Petar Veličković (Deepmind) - Categories, Graphs, Reasoning [NEURIPS22 UNPLUGGED]
Machine Learning Street Talk (MLST)
Neural Networks Are Not Turing Machines but They Can Extrapolate
The key is to find the right sweet spot between full universal approximator mlps and algorithms on the other side. latent space where this gnn can best do its thing that really is software 2.0 but i wanted to ask you about the computational limitation because you said something interesting about representing you know infinite objects with with a finite memory so neural networks are not Turing machines but they can extrapolate of courseYeah what's the realistic limitation let's say you're trying to learn an algorithm how far can you go with the neural network? There are cases where you can go very far we do have theory that is very robust about this and i think it's theory that is actually quite easily understandable
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