←back to thread

760 points MindBreaker2605 | 6 comments | | HN request time: 1.852s | source | bottom
1. gregjw ◴[] No.45897653[source]
Interesting he isn't just working with Feifei Li if he's really interested in 'world models'.
replies(3): >>45897794 #>>45898810 #>>45905955 #
2. muragekibicho ◴[] No.45897794[source]
Exactly where my mind turned. It's interesting how the AI OG's (Feifei and Cunn) think world models are the way forward.
3. dauertewigkeit ◴[] No.45898810[source]
Correct me if I'm wrong but LeCun is focused on learning from video, whereas Fei-Fei Li is doing robotic simulations. Also I think Fei-Fei Li's approach is still using transformers and not buying into JEPA.
replies(1): >>45903109 #
4. Legend2440 ◴[] No.45903109[source]
JEPA is not an alternative to transformers, it is built out of transformers.
replies(1): >>45905511 #
5. HarHarVeryFunny ◴[] No.45905511{3}[source]
JEPA is trying to learn real world regularities by predicting internal perceptual representations (using a transformer).

It seems a good direction to go in, but tbh he doesn't seem to have taken it very far, and it is the transformer doing all the heavy lifting.

6. intalentive ◴[] No.45905955[source]
LeCun is not interested in generative models which predict in pixel space. He wants "energy based" models which predict in the representation space.