3 ms·
Self-Supervised Learning and Foundation Models
- aliqot 4y agoI preferred the alternative title: "What happens when you pour gasoline on your own bias"
- rgabrielsson 4y agohaha future will tell! :D
- nadermx 4y agoIt's nice MIT offers lectures like this and the gilbert strands lectures. Any other ML/Math related lectures recomended?
- mdp2021 4y agoThe classic 6.034 of [Classic] Artificial Intelligence from the late Prof. Patrick Winston is "mandatory".
- rgabrielsson 4y agoOne trend I'm seeing: I think the math that is becoming more useful, if huge foundation models are taking over, is around comparing and understanding their embedding/latent spaces. Geometry, topology, optimal transport, alignment -- in favor of statistics and probability theory