Demis confirmed DeepMind hasn’t cracked continual learning.
Yuchen Jin
X profili@Yuchenj_UW92 B takipçi
75.633 görüntüleme
Demis confirmed DeepMind hasn’t cracked continual learning.
There were rumors on X that Google DeepMind had solved it. Not true.
They’re experimenting with blending AlphaZero with large foundation models to see if it can scale and generalize to the real world, not just games.
Yorumlar
Yorumlar X (Twitter) üzerindeki orijinal gönderiden alınmıştır
Henüz yorum yok
