@mengjiao_yang Profile picture

Sherry Yang

@mengjiao_yang

Research Scientist @GoogleDeepMind | PhD Student @UCBerkeley. Previously M.Eng. / B.S. @MIT.

Similar User
Abhishek Gupta photo

@abhishekunique7

Shuran Song photo

@SongShuran

Jakob Foerster photo

@j_foerst

Igor Mordatch photo

@IMordatch

Dorsa Sadigh photo

@DorsaSadigh

Ted Xiao photo

@xiao_ted

Nathan Lambert photo

@natolambert

Rishabh Agarwal photo

@agarwl_

Hao Liu photo

@haoliuhl

Nan Jiang photo

@nanjiang_cs

Ilya Kostrikov photo

@ikostrikov

Andy Zeng photo

@andyzeng_

Ruiqi Gao photo

@RuiqiGao

Yilun Du photo

@du_yilun

Qinqing Zheng photo

@qqyuzu

Sherry Yang Reposted

This was a really fun collaboration between experts in materials science, diffusion, and foundation models. Connecting material generation with foundation models (although still loosely in this work) has so many exciting implications!

Checkout Generative Hierarchical Materials Search (GenMS) – a framework for generating crystal structures from natural language. Website: generative-materials.github.io Paper: arxiv.org/abs/2409.06762



Source for this figure: arxiv.org/abs/2205.10816. Procedure Cloning is a simple but powerful idea: Teach the model not just what action to take but also the procedure for how to find this action. Original Tweet: x.com/mengjiao_yang/…

My Bet: Strawberry is algorithm distillation/procedural cloning. Everyone right now is coming up with ways to distill System 2 into System 1, but that will always be limited. We need to train the model to run the algorithms, not just outputs (and post-train with RL of course).

Tweet Image 1


Looking forward to presenting the following papers @icmlconf: - Position paper on Video Generation for Decision Making arxiv.org/abs/2402.17139 (Tue 1:30 - 3pm #2613). - Code as Reward for real-world RL with VLMs arxiv.org/abs/2402.04764 (Thur 1:30-3pm #1115).


Loading...

Something went wrong.


Something went wrong.