/Tech7h ago

ICML awards Outstanding Paper to diffusion language model study, marking three consecutive years of diffusion wins

The paper argues that arbitrary ordering hinders diffusion-based generation.

699125020.9K

#838

Original post

Volodymyr Kuleshov 🇺🇦@volokuleshov#838inTech

Super exciting that another ICML best paper award goes to work on diffusion language models.

That's now three ICML's in a row where one of the best papers is on diffusion for language: SEDD (2024), Train for the Worst and Plan for the Best (2025), and now the Flexibility Trap (2026).

ICML Conference@icmlconf

🏆Outstanding Paper is awarded to the following research papers:

The Flexibility Trap: Rethinking the Value of Arbitrary Order in Diffusion Language Models

High-Accuracy Sampling for Diffusion Models and Log-Concave Distributions

2:16 AM · Jul 6, 2026 · 4.5K Views

Sentiment

Many users congratulated ICML award winners for diffusion model studies because they viewed the papers as impressive and felt personally honored by the recognition.

Pos

100.0%

Neg

0.0%

6 comments with sentiment.

Cluster Engagement

Digg Deeper

No Digg Deeper questions have been answered for this story yet.

Posts from X

Most Activity

VIEWS37.8KBOOKMARKS58LIKES76RETWEETS10REPLIES4

ICML Conference@icmlconf

🏆Five research papers are recognized as Honorable Mention for the Outstanding Paper award (1/2):

The Obfuscation Atlas: Mapping Where Honesty Emerges in RLVR with Deception Probes

Motion Attribution for Video Generation

How much can language models memorize?

(cont)...

17h37.8K7658

ICML Conference@icmlconf

🏆Finally, the Test of Time Award:

Asynchronous Methods for Deep Reinforcement Learning

Originally published at #ICML2016

17h7.6K3410

ICML Conference@icmlconf

🏆Five research papers are recognized as Honorable Mention for the Outstanding Paper award (2/2):

(cont) ...

A Random Matrix Perspective on the Consistency of Diffusion Models

To Grok Grokking: Provable Grokking in Ridge Regression

17h4.4K256

ICML Conference@icmlconf

🏆Outstanding Paper is also awarded to the following position paper:

Position: The Alignment Community is Unintentionally Building a Censor’s Toolkit

17h5.5K285

ICML Conference@icmlconf

🏆One position paper is recognized as Honorable Mention for the Outstanding Paper award:

Position: AI/ML Deepfake Research is Misaligned with AI Generated Non-Consensual Intimate Imagery (AIG-NCII)

17h3.3K114

ICML Conference@icmlconf

Again, please check out the full blog post for more details on all winners and the selection process. Congrats to all the winners! 🏆🏆🏆

https://blog.icml.cc/?p=1347

17h3.7K113

Jonathan Lorraine@jonLorraine9

So honored that MOTIVE was named an ICML 2026 Outstanding Paper Honorable Mention! 🎉

Motion Attribution for Video Generation 👇 http://research.nvidia.com/labs/sil/projects/MOTIVE/

TL;DR: We propose Motive, a scalable, motion-centric data attribution framework for video generation to identify which training clips improve or degrade motion dynamics, enabling curation and more.

Huge credit to Xindi Wu (@cindy_x_wu), who led this, and to the whole team.

Congrats and thanks to @paschalidoud_1 @JunGao33210520 Antonio Torralba @lealtaixe @orussakovsky @FidlerSanja, with support from @NVIDIAAI @PrincetonCS @MIT_CSAIL @UMichCSE @UofTCompSci @VectorInst 💗 #ICML2026

14h1.8K82

Brokoslaw Laschowski@DrLaschowski

@icmlconf Congrats Volodymyr and the entire @GoogleDeepMind team

15h1462

Aditya Kusupati@adityakusupati

@icmlconf Congratulations Jack! @jxmnop.

Impressive paper.

15h4441

Aaron Gokaslan@SkyLi0n

@icmlconf Congrats @jxmnop !

15h2561

Dr. Phil Hackemann@PhilHackemann

@icmlconf 🙏🏼 Thank you so much for recognizing our work! @sarahba1010 and I feel very honored!

If you want to learn more about our position paper, visit our project website (there you can also find the dates for our oral presentation and poster session): https://s-ball-10.github.io/censors-toolkit/

14h292

Sarah Ball@sarahba1010

@icmlconf 🥳 Thanks so much! This is incredible 🤩 Looking forward to the conference! Let‘s meet at our oral presentation and poster session 🤝 Learn more about our paper here: https://s-ball-10.github.io/censors-toolkit/

13h10