Picture

Po-Wei Wang @_powei

Applied scientist at Pinterest & PhD from @SCSatCMU. Optimization, ML, graph, and GPU accelerations. ✈️ RecSys 2024 @ Bari. powei.tw Palo Alto, CA Joined October 2010
  • Tweets
    419
  • Followers
    541
  • Following
    724
  • Likes
    1K
Po-Wei Wang retweeted
ZeyuanAllenZhu Profile Picture

Zeyuan Allen-Zhu, Sc.D. @ZeyuanAllenZhu

5 months ago

(1/8)🍎A Galileo moment for LLM design🍎 As Pisa Tower experiment sparked modern physics, our controlled synthetic pretraining playground reveals LLM architectures' true limits. A turning point that might divide LLM research into "before" and "after." physics.allen-zhu.com/part-4-archite…

Po-Wei Wang retweeted
keenanisalive Profile Picture

Keenan Crane @keenanisalive

6 months ago

Here's a nice "proof without words": The sum of the squares of several positive values can never be bigger than the square of their sum. This picture helps make sense of how ℓ₁ and ℓ₂ norms regularize and sparsify solutions (resp.). [1/n]

Po-Wei Wang retweeted
hankyang94 Profile Picture

Heng Yang @hankyang94

a year ago

I had a blast talking about "The Power of (Modern) Convex Optimization in Nonconvex Perception, Control, and Learning" at Northeastern University. I highlighted two recent strands of research from my group: (1) Convex relaxation + large-scale numerical optimization + structure…

Po-Wei Wang retweeted
polynoamial Profile Picture

Noam Brown @polynoamial

a year ago

Today, I’m excited to share with you all the fruit of our effort at @OpenAI to create AI models capable of truly general reasoning: OpenAI's new o1 model series! (aka 🍓) Let me explain 🧵 1/

Po-Wei Wang retweeted
ZeyuanAllenZhu Profile Picture

Zeyuan Allen-Zhu, Sc.D. @ZeyuanAllenZhu

a year ago

(1/7) Physics of LM, Part 2.2 with 8 results on "LLM how to learn from mistakes" now on arxiv: arxiv.org/abs/2408.16293. We explore the possibility to enable models to correct errors immediately after they are made (no multi-round prompting). Check out the slides in this thread.

Po-Wei Wang retweeted
tri_dao Profile Picture

Tri Dao @tri_dao

a year ago

Hybrid Mamba / Transformer architecture works really well, especially for long context and fast inference

AI21Labs Profile Picture

AI21 Labs @AI21Labs

a year ago

Hybrid Mamba / Transformer architecture works really well, especially for long context and fast inference

Po-Wei Wang retweeted
ZeyuanAllenZhu Profile Picture

Zeyuan Allen-Zhu, Sc.D. @ZeyuanAllenZhu

a year ago

(1/2) Many asked for Part 2.2 and I'm sorry for the delay. Our author Zicheng Xu has been unexpectedly laid off. He has my strongest endorsement (see next post). If interested in this project or hiring him, contact him at [email protected] (remove the capital 'B')

ZeyuanAllenZhu Profile Picture

Zeyuan Allen-Zhu, Sc.D. @ZeyuanAllenZhu

a year ago

(1/2) Many asked for Part 2.2 and I'm sorry for the delay. Our author Zicheng Xu has been unexpectedly laid off. He has my strongest endorsement (see next post). If interested in this project or hiring him, contact him at [email protected] (remove the capital 'B')

Po-Wei Wang retweeted
DBertsekas Profile Picture

Dimitri Bertsekas @DBertsekas

2 years ago

Posted Youtube video youtube.com/watch?v=kYP-mK… of "A 60-year journey in #convex #optimization," a lecture on the history and the evolution of the subject. From a 2009 MIT event honoring S Mitter. Includes a review of the MCMC abstract duality framework.

Po-Wei Wang retweeted
Jeande_d Profile Picture

Jean de Nyandwi @Jeande_d

2 years ago

Large Language Models and Transformers - Simons Institute Excellent lectures on ongoing revolution in transformers and large language models. Topics range from recent publications, frontier/foundational works, and all things LLMs. Features people who are at the forefront of AI…

Po-Wei Wang retweeted
PinterestEng Profile Picture

Pinterest Engineering @PinterestEng

2 years ago

Day Two: #KDD2023 Meet the Pinterest team at our sessions today! 📌

Po-Wei Wang retweeted
khainb_ml Profile Picture

Khai Nguyen @khainb_ml

2 years ago

I also would like to thank @brandondamos for his amazing tutorial on amortized optimization (AO) arxiv.org/abs/2202.00665 which helps me a lot when applying AO to SW. I strongly recommend this if you want to know a comprehensive usage of AO in machine learning. (12/12)

Po-Wei Wang retweeted
mathOCb Profile Picture

arXiv math.OC Optimization and Control @mathOCb

2 years ago

Christoph Buchheim: Efficient KKT reformulations for bilevel linear programming arxiv.org/abs/2307.06639 arxiv.org/pdf/2307.06639

Po-Wei Wang retweeted
KaiyuYang4 Profile Picture

Kaiyu Yang @KaiyuYang4

2 years ago

Can LLMs generate mathematical proofs that can be rigorously checked? We release LeanDojo (leandojo.org): an open-source playground consisting of toolkits, benchmarks, and models for LLMs to prove formal theorems in the Lean proof assistant. Key features: - Tools for…

Po-Wei Wang retweeted
Tim_Dettmers Profile Picture

Tim Dettmers @Tim_Dettmers

2 years ago

@karpathy Super excited to push this even further: - Next week: bitsandbytes 4-bit closed beta that allows you to finetune 30B/65B LLaMA models on a single 24/48 GB GPU (no degradation vs full fine-tuning in 16-bit) - Two weeks: Full release of code, paper, and a collection of 65B models

Po-Wei Wang retweeted
BingbinL Profile Picture

Bingbin Liu @BingbinL

2 years ago

(1/7) Couldn’t make it to Kigali but would like to share our ICLR oral: Transformers solve multi-step algorithmic reasoning with surprisingly few layers – why? We study this via automata and show that o(T)-layer Transformers (aka. shortcuts) can represent T steps of reasoning.

Po-Wei Wang retweeted
PinterestEng Profile Picture

Pinterest Engineering @PinterestEng

2 years ago

Tomorrow we're hosting our next Pinterest Labs Talk Series event with Cho-Jui Hsieh, Associate Professor of Computer Science at UCLA. Join us there 📌 bit.ly/3GDO8if

Po-Wei Wang retweeted
LinusHNielsen Profile Picture

Linus Härenstam-Nielsen @LinusHNielsen

2 years ago

Happy to announce that our paper "Semidefinite Relaxations for Robust Multiview Triangulation" was accepted at #CVPR2023! paper: arxiv.org/abs/2301.11431 code: github.com/Linusnie/robus… co-authors: @NiclasZeller and Daniel Cremers

Po-Wei Wang retweeted
_chris_lu_ Profile Picture

Chris Lu @_chris_lu_

3 years ago

1/ 🚀 Presenting PureJaxRL: A game-changing approach to Deep Reinforcement Learning! We achieve over 4000x training speedups in RL by vectorizing agent training on GPUs with concise, accessible code. Blog post: chrislu.page/blog/meta-disc… 🧵

Trends for United States

Nicki
486 B posts
The Twitter Give
346 B posts
Happy New Month
141 B posts
Independence Day
49,9 B posts
#NigeriaAt65
5.429 posts
#governmentshutdown2025
2.196 posts
Aces
35,8 B posts
Kelsey Mitchell
8.084 posts
Fever
46,6 B posts
Dodgers
93,8 B posts
Happy Uptober
8.297 posts
Sora 2
9.016 posts
Odyssey Sims
3.873 posts
#DWTS
40,8 B posts
Boone
28,3 B posts
WNBA
98,1 B posts
Kulture
45 B posts
Lauren
62,2 B posts
$WEED
7.777 posts
Spooky Month
11,8 B posts

You might like

Professor and Head of Machine Learning Department at @CarnegieMellon. Board member @OpenAI and @Qualcomm. Chief Technical Advisor @GraySwanAI.
Zico Kolter
@zicokolter
24K Followers
Assistant professor at Stanford; Co-founder of Voyage AI (https://t.co/wpIITHLgF0) ; Working on ML, DL, RL, LLMs, and their theory.
Tengyu Ma
@tengyuma
37K Followers
Assistant Professor @MIT | Co-founder & Chair @ClimateChangeAI | MIT #TR35, #TIME100AI | she/they
Priya L. Donti
@priyald17
5K Followers
Member of Technical Staff @thinkymachines | prev CS PhD @CSDatCMU
Mingjie Sun
@_mingjiesun
767 Followers
machine learning and optimization @PrincetonCS & Google DeepMind Princeton, dad^3
Elad Hazan
@HazanPrinceton
14K Followers
Assistant Professor MIT @medialab @MITEECS @nlp_mit || PhD from CMU @mldcmu @LTIatCMU || Foundations of multisensory AI to enhance the human experience.
Paul Liang
@pliang279
8K Followers
optimization. machine learning. uc berkeley. I blog at https://t.co/fkJujOPsJb The world won't end.
Ben Recht
@beenwrekt
32K Followers
Research fellow at Flatiron Institute, working on understanding optimization in deep learning. Previously: PhD in machine learning at Carnegie Mellon.
Jeremy Cohen
@deepcohen
5K Followers
Chief Scientist @LumaLabsAI. #LumaDreamMachine
Jiaming Song
@baaadas
9K Followers
Research Scientist. FAIR NY, Meta. I build simplified abstractions of the world through the lens of dynamics and flows.
Ricky T. Q. Chen
@RickyTQChen
6K Followers
Research @OpenAI. Trained models for GPT5 Thinking / Mini; Contributor to gpt-oss, o3-mini, o1. Previously @SFResearch, PhD @Stanford.
Yu Bai
@yubai01
6K Followers
Researcher in machine learning
Francis Bach
@BachFrancis
18K Followers
Assistant Professor @uwcse. Postdoc @the_IAS. PhD in machine learning @mldcmu.
Simon Shaolei Du
@SimonShaoleiDu
9K Followers
Professor at ETH Zurich
Andreas Krause
@arkrause
6K Followers
Assistant professor @UMichCSE @UMich; previously @SimonsInstitute @UCBerkeley @Princeton @Tsinghua_Uni. Theoretical and scientific foundations of deep learning.
Wei Hu
@weihu_
2K Followers