Picture

Yunhao (Robin) Tang @robinphysics

Interested in RL. Science @MistralAI. Prev Llama post-training @AIatMeta, Gemini post-training and deep RL research @Deepmind, PhD @Columbia robintyh1.github.io Joined November 2018
  • Tweets
    128
  • Followers
    1K
  • Following
    729
  • Likes
    1K
Yunhao (Robin) Tang retweeted
MistralAI Profile Picture

Mistral AI @MistralAI

3 months ago

Announcing Magistral, our first reasoning model designed to excel in domain-specific, transparent, and multilingual reasoning.

Yunhao (Robin) Tang retweeted
ZacKenton1 Profile Picture

Zac Kenton @ZacKenton1

a year ago

Eventually, humans will need to supervise superhuman AI - but how? Can we study it now? We don't have superhuman AI, but we do have LLMs. We study protocols where a weaker LLM uses stronger ones to find better answers than it knows itself. Does this work? It’s complicated: 🧵👇

robinphysics Profile Picture

Yunhao (Robin) Tang @robinphysics

a year ago

Thanks @_akhaliq for promoting our work! Unlike regular RL where golden r(s,a) are available and online is generally deemed better than offline, in RLHF this is less clear. Complementary to some concurrent work, we investigate causes to the perf gap between online vs. offline.

_akhaliq Profile Picture

AK @_akhaliq

a year ago

Thanks @_akhaliq for promoting our work! Unlike regular RL where golden r(s,a) are available and online is generally deemed better than offline, in RLHF this is less clear. Complementary to some concurrent work, we investigate causes to the perf gap between online vs. offline.

Yunhao (Robin) Tang retweeted
misovalko Profile Picture

Michal Valko @misovalko

2 years ago

Fast-forward ⏩ alignment research from @GoogleDeepMind ! Our latest results enhance alignment outcomes in Large Language Models (LLMs). Presenting NashLLM!

robinphysics Profile Picture

Yunhao (Robin) Tang @robinphysics

2 years ago

Interested in how **non-contrastive representation learning for RL** is magically equivalent to **gradient-based PCA/SVD on the transition matrix** and hence won't collapse and capture spectral info about the transition? Come talk to us at #ICML2023 Hall 1 #308 at 1:30pm

robinphysics Profile Picture

Yunhao (Robin) Tang @robinphysics

2 years ago

Interested in how **non-contrastive representation learning for RL** is magically equivalent to **gradient-based PCA/SVD on the transition matrix** and hence won't collapse and capture spectral info about the transition? Come talk to us at #ICML2023 Hall 1 #308 at 1:30pm

Yunhao (Robin) Tang retweeted
wwdabney Profile Picture

Will Dabney @wwdabney

2 years ago

Even if all you want is a value function, using quantile TD (QTD) can give a better estimate than standard TD. Today at #ICML2023, Mark Rowland presents our latest work on distributional RL in collaboration with @robinphysics, @clarelyle, Remi Munos, @marcgbellemare #809 @ 2pm

robinphysics Profile Picture

Yunhao (Robin) Tang @robinphysics

2 years ago

Interested in how non-contrastive representation learning works in RL? We show (1) Why representations do not collapses (2) How it relates to gradient PCA / SVD of transition matrix Understanding Self-Predictive Learning for RL #ICML2023 @GoogleDeepMind arxiv.org/pdf/2212.03319

Trends for United States

Chiefs
121 B posts
#SmackDown
58,4 B posts
Worthy
45,5 B posts
Mahomes
25,5 B posts
Herbert
15,2 B posts
Kelce
23,6 B posts
#KCvsLAC
2.806 posts
Sami
31,6 B posts
Hollywood Brown
2.547 posts
Rich Eisen
Brazil
125 B posts
Ryan Clark
10,2 B posts
Mr. Beast
6.996 posts
#TheLastDriveIn
1.167 posts
Quentin Johnston
2.182 posts
Brock
17,8 B posts
Ladd
3.106 posts
Tylenol
32,5 B posts
#BoltUp
3.480 posts
Louisville
4.825 posts

You might like

Reinforcement Learner, Adjunct Prof at McGill. Ex MSL Meta, DeepMind, Brain, Mila, IIT Bombay. NeurIPS Best Paper
Rishabh Agarwal
@agarwl_
17K Followers
Assoc Prof in ML @UniofOxford @StAnnesCollege @FLAIR_Ox/ RS @MetaAI, 2x dad. Ex: (A)PM @Google, DivStrat @GS, ex intern: @GoogleDeepmind, @GoogleBrain, @OpenAI
Jakob Foerster
@j_foerst
21K Followers
CSO & co-founder, Reliant AI. Ex RL research lead at Google Brain, DeepMind. Known for Atari 2600 RL benchmark, Distributional RL (MIT Press 2023).
Marc G. Bellemare
@marcgbellemare
16K Followers
Assistant Professor at University of Washington. I like robots, and reinforcement learning. Previously: post-doc at MIT, PhD at Berkeley
Abhishek Gupta
@abhishekunique7
9K Followers
machine learning researcher, with focus on reinforcement learning. assoc prof @ uiuc cs. Course on RL theory (w/ videos): https://t.co/vqVKwY4RJE
Nan Jiang
@nanjiang_cs
10K Followers
This is the site where I talk about the attacks on science and immigration. Science is on the other site. Lab website: https://t.co/vrtbcqRyRn
Eugene Vinitsky (@RLC...
@EugeneVinitsky
20K Followers
machine learning and optimization @PrincetonCS & Google DeepMind Princeton, dad^3
Elad Hazan
@HazanPrinceton
14K Followers
Associate professor, Computer Science. Stanford. Stanford's Human Centered AI (HAI) Institute. Opinions expressed are my own.
Emma Brunskill
@EmmaBrunskill
8K Followers
VP of Research @googledeepmind.
raia hadsell
@RaiaHadsell
9K Followers
ML theory nerd & AI non-enthusiast. thinking a lot about online learning these days! BTW you should go find me on another website where i post more actively
Gergely Neu
@neu_rips
11K Followers
Assistant Professor @uwcse. Postdoc @the_IAS. PhD in machine learning @mldcmu.
Simon Shaolei Du
@SimonShaoleiDu
9K Followers
Associate Professor @NorthwesternU | PhD @Princeton | studying Reinforcement Learning
Zhaoran Wang
@zhaoran_wang
4K Followers
Building something new · Chief Models Officer @ Stealth Startup & Inria & MVA - Ex: Llama @AIatMeta Gemini and BYOL @GoogleDeepMind
Michal Valko
@misovalko
8K Followers
Csaba Szepesvari
@CsabaSzepesvari
11K Followers
石宇歌 · @GoogleDeepMind, Genie 1, 2 🧞‍♀️ , Veo 3 📽️
Yuge Shi (Jimmy)
@YugeTen
5K Followers