I'm hiring 2 PhD students & 1 postdoc @GeorgiaTech for Fall'26 Motivated students plz consider us, especially those in * ML+Quantum * DeepLearning+Optimization -PhD: see https://t.co/h4anjm6b8j -Postdoc: see https://t.co/548XVaahx3 & https://t.co/4ahNE7OOwV Retweet appreciated

Drowning in the sea of Discrete Diffusion papers? 🌊 We got you. Join our Reading Group! From theory → empirics, and language → molecules — we’ll decode the chaos together 💫 Join the cult—uh, I mean community 😇 👉 Google Group:  https://t.co/kV9efqBBTu (1 / 2)

+1 way to apply for postdoc in ML theory @GeorgiaTech This one https://t.co/N4UKTxxgmJ is for 2 yrs. You can work with both me & other faculty in Math, CS & ISyE. Other positions (3 yr): AI4Science: https://t.co/548XVaaPmB regular: https://t.co/4ahNE7Pmmt https://t.co/L06RwiEJU5

A short-and-sweet guide to developing research questions published in @Nature. Lots of good advice here 👍

https://t.co/ockRvJhTL4

Success of RL post-training hinges on the quality of generated rollouts, but high-reward targets are sparsely scattered in the vast state space, hindering the effectiveness of reward optimization💫. 🧩Solution? 💡𝐒𝐞𝐚𝐫𝐜𝐡-𝐭𝐲𝐩𝐞 Inference-time Scaling +

Super happy to share our new work on “Tree Search Guided Trajectory-Aware Fine-Tuning for Discrete Diffusion” or TR2-D2! 🤖🌳 Inspired by the incredible success of off-policy reinforcement learning (RL), TR2-D2 introduces a general framework that combines off-policy RL with tree

Proud of my junior collaborators Kijung Jeon Yuchen @YuchenZhu_ZYC Wei @WeiGuo01 Jaemoo @jaemoo51133 Avrajit @GhoshAvrajit Lianghe Shi Yinuo @Yinuo_Ren Haoxuan @haoxuan_steve_c - 6 joint #NeurIPS2025 main track paper! Lucky to have you Wanna join us? Will post recruit info soon.

Wrapped up Stanford CS336 (Language Models from Scratch), taught with an amazing team @tatsu_hashimoto @marcelroed @neilbband @rckpudi. Researchers are becoming detached from the technical details of how LMs work. In CS336, we try to fix that by having students build everything:

Wei Guo @WeiGuo01 also wrote some nice posts with more details. Please check them out if interested!

Kudos to my amazing collaborators @YuchenZhu_ZYC @WeiGuo01 @jaemoo51133 @guanhorng_liu @YongxinChen1 ! See Yuchen Zhu @YuchenZhu_ZYC's posts for more details!

Sampling is hard b/c target distribution can be high-dim with many modes. ML can help, even when state space is discrete (thus non-differentiable)! https://t.co/vNtqYOzQc5 constructs a strong sampler by fine tuning a discrete diffusion model via stochastic optimal control / RL.

On the coming Tuesday (Aug 26th), we will have @YuchenZhu_ZYC talking about “Beyond Euclidean data: Lie group and multimodal diffusion models"🚀, from 5pm to 6pm (UK time). Join us via zoom: https://t.co/N1C3UFukxd See more information below 👇

By popular demand, #NeurIPS2025 Workshop on "Dynamics at the Frontiers of Optimization, Sampling, and Games" (DynaFront) has its submission deadline extended to August 29 (AoE). Please submit high quality work at

Georgia Tech AI4Science Center is soft launched, and I'm excited to be an Associate Director. https://t.co/jOuS3g3J5j Collaboration+Participation of all kinds are welcomed. Please get in touch! Thanks to @gtsciences for supports. Retweets appreciated! @GeorgiaTech #AI4Science

There is still time to become a reviewer for @NeurIPSConf 2025 Workshop DynaFront (Dynamics at the Frontiers of Optimization, Sampling, and Games) Plz register at https://t.co/h06XhVbnMB You can also submit high quality manuscripts till Aug 22. https://t.co/lUl7uKHGR7

Interested in some foundation aspects? Waiting or unhappy about NeurIPS reviews? Plz consider NeurIPS workshop DynaFront: Dynamics at the Frontiers of Optimization, Sampling, and Games https://t.co/lUl7uKHGR7 @yuejiec @Andrea__M @btreetaiji @T_Chavdarova ++ Sponsor appreciated!

I will present * accelerated manifold optimization, in LA (ICCOPT) Wed 7/23 * fine tuning of diffusion model and stochastic optimal control for sampling, in Montreal (SIAM) Tue 7/29 * fast sampling under nonconvex constraints, in Chicago (MCM) Thu 7/31 Love to chat and learn!

If still around #ICML2025, plz consider checking out my collaborator @qu_1006 's Oral in the MemFM Workshop, 11am Sat West Meeting Room 223-224, on A Closer Look at Model Collapse (in diffusion model): From a Generalization-to-Memorization Perspective

What if AI isn’t about building solo geniuses, but designing social systems? Michael Jordan advocates blending ML, economics, and uncertainty management to prioritize social welfare over mere prediction. A must-read rethink. https://t.co/HUJh97pq5N