Scaling Scaling Laws with Board Games · HackerPapers discussions

Scaling Scaling Laws with Board Games — mentioned by espadrine

Fri, 10 Mar 2023 08:28:17 GMT

Model sizes are adjusted to reach a given quality. Many papers justified empirically the use of more neurons allocated to language than to visual understanding, like Imagen[0] (figure 4.a). as it yields better results. Debatably, it is also true in the human brain. To me, it could highlight that there is more entropy in language than in images. After all, games have a similar property: some games, like noughts and crosses, or checkers, can have the same model architecture reach optimal play with smaller sizes than are necessary for chess and go[1]. It is certainly true that language has a lot…

Paper and discussion history. A mention is not an endorsement.

Scaling Scaling Laws with Board Games — mentioned by gwern

Fri, 19 Aug 2022 22:46:48 GMT

> it isn't immediately clear that current RL algorithms will neatly scale with data & parameter count It may not be immediately clear, but it is nevertheless unfortunately clear from RL papers which provide adequate sample-size or compute ranges that RL appears to follow scaling laws (just like everywhere else anyone bothers to test). Yeah, they just get better the same way that regular ol' self-supervised or supervised Transformers do. Sorry if you were counting on 'RL doesn't work' for safety or anything. If you don't believe the basic existence proofs of things like OA5 or AlphaStar, which…

Paper and discussion history. A mention is not an endorsement.

Scaling Scaling Laws with Board Games — mentioned by gwern

Thu, 28 Apr 2022 18:29:25 GMT

> Scaling laws for RL agents are discovered, similar to language models. Still waiting for a followup to https://arxiv.org/abs/2104.03113 ...

Paper and discussion history. A mention is not an endorsement.