Grigoris Chrysos
@Grigoris_c
Assistant Professor at @UWMadison. Interested in representation learning and trustworthy ML. https://grigoris.ece.wisc.edu/
You might like
Fun activity with agent design ⬇️
Excited to announce that we're hosting a language-model agent design challenge for playing StarCraft 2, Street Fighter 3, Super Mario, Pokémon, and 2048! The total prize pool is $20K! - $10K for the small language model track - $10K for the LLM track. Thanks to generous…
Kudos to UCSD for the report; well articulated thoughts on preparing students
A very well-argued article about the UCSD remedial math report that drew a lot of attention. It makes two great points about schools in the US, and the UC system. It should be noted that the UCSD leadership actually commissioned this report/committee; and the faculty diligently…
Transparency for the win. Despite the negative publicity, @iclr_conf format of open reviews exposes the problem and enables PCs to suggest solutions (🙏). Can we propose solutions instead of simply complaining for our scores at the conference?
We are aware of low-quality and LLM-generated reviews and are currently deliberating on appropriate courses of action. For now, authors who receive very poor quality or LLM-generated reviews should flag them to their ACs. We appreciate the community's efforts in reporting these!
Please share widely. The UW–Madison ECE Department is recruiting at the assistant, associate, or full professor in foundational AI and machine learning tinyurl.com/4kwh8mvm
Positive steps towards reducing the influx of papers without technical results.
Much needed step from arXiv for quality control and preventing blog post & survey dumping: blog.arxiv.org/2025/10/31/att…
Thank you for preparing this great resource. Interested in diffusion models? Start below from experts in the topic
Tired of chasing references across dozens of papers? This monograph distills it all: the principles, intuition, and math behind diffusion models. Thrilled to share!
Working on fine-tuning or post-training? You can submit to a special issue in the highly influential journal of IJCV. 📘 Link: ijcvsi25pt.github.io 📅 Submission timeline: anytime before July 15
🚀 We’re organizing @IJCV special issue on “Post-Training in LLMs for Computer Vision”! With guest editors @Grigoris_c @vidal_rene Dacheng Tao (NTU) and Philip Torr (Oxford) 📘 Learn more & submit your work: ijcvsi25pt.github.io 📅 Submission ddl: July 15, 2026
Direct call for revisiting all of the current frontier-lab safety approaches 🔽
One cannot show that turbojets are safe before actually building turbojets and carefully refining them for reliability. The same goes for AI.
Applying for faculty positions? Join us in Madison
UW–Madison ECE is seeking a faculty member in foundational AI/ML (all ranks). Areas: trustworthy AI, scalable architectures, optimization, AI-hardware co-design, and applications in signals, sensing, comms & energy. More info: jobs.wisc.edu/jobs/assistant… @UWMadEngr
TMLR is looking for additional reviewers and action editors! Please sign up to keep the initiative strong and vibrant.
As Transactions on Machine Learning Research (TMLR) grows in number of submissions, we are looking for more reviewers and action editors. Please sign up! Only one paper to review at a time and <= 6 per year, reviewers report greater satisfaction than reviewing for conferences!
Small models as the new frontier and why this may be academia's LLM moment Academia should reject the nihilism of "scale is all you need", i.e, that meaningful research requires frontier scale compute. This mindset hurts basic research and what we can contribute to machine…
I will be admitting a select few highly motivated PhD students to join the CausalML Lab @JHUCompSci in Fall 2026. If you are interested in causal inference, generative models, and ML, apply to the JHU CS PhD program and list me as a potential advisor.
Excited to learn more from your creative approach and spins on the years ahead! Congratulations on your well deserved tenure! 🍻
Happy to share that I got tenured last month! While every phase in life is special, this one feels a bit more meaningful, and it made me reflect on the past 15+ years in academia. I'd like to thank @UWMadison and @UWMadisonECE for tremendous support throughout the past six…
Great video for a smooth introduction to text diffusion models
Diffusion LLMs are promising ways to overcome the limitations of autoregressive LLMs. Less error propagation, easier to control, and faster to sample! But how do Diffusion LLMs actually work? 🤔 Let's explore some ideas on this fascinating topic! youtu.be/8BTOoc0yDVA
⬇️ Very good advice
Instead of complaining that peer review is dead, take a positive step to improve it today. The reviewers are not aliens, they are us! - Revise your review and make it clear. Identify the crucial points that impacted your score negatively and positively. - If the paper is…
Indeed, why is code not mandatory? That could solve various problems.
The NeurIPS paper checklist corroborates the bureaucratic theory of statistics. argmin.net/p/standard-err…
🚨 Panel on "how are theoretical tools useful in vision?" with an amazing list of panelists: @CevherLIONS @orussakovsky @vidal_rene Open to your questions, the more ambitious the better. In @CVPR : Room 107 A at 12 🎸.
Sparsity but with increased expertise across tens of thousands of combinations? Check out the details from James ⬇️
Sparse MLPs/dictionaries learn interpretable features in LLMs, yet provide poor layer reconstruction. Mixture of Decoders (MxDs) expand dense layers into sparsely activating sublayers instead, for a more faithful decomposition! 📝 arxiv.org/abs/2505.21364 [1/7]
![jamesaoldfield's tweet image. Sparse MLPs/dictionaries learn interpretable features in LLMs, yet provide poor layer reconstruction.
Mixture of Decoders (MxDs) expand dense layers into sparsely activating sublayers instead, for a more faithful decomposition!
📝 arxiv.org/abs/2505.21364
[1/7]](https://pbs.twimg.com/media/Gsd1yNAXgAA0qdJ.png)
![jamesaoldfield's tweet image. Sparse MLPs/dictionaries learn interpretable features in LLMs, yet provide poor layer reconstruction.
Mixture of Decoders (MxDs) expand dense layers into sparsely activating sublayers instead, for a more faithful decomposition!
📝 arxiv.org/abs/2505.21364
[1/7]](https://pbs.twimg.com/media/Gsd18AkW0AA0huF.jpg)
United States Trends
- 1. #GMMTV2026 467K posts
- 2. MILKLOVE BORN TO SHINE 55.3K posts
- 3. #WWERaw 76.6K posts
- 4. Panthers 37.7K posts
- 5. Purdy 28.3K posts
- 6. Finch 14.4K posts
- 7. AI Alert 7,990 posts
- 8. TOP CALL 9,203 posts
- 9. Moe Odum N/A
- 10. Bryce 21.2K posts
- 11. Timberwolves 3,881 posts
- 12. Keegan Murray 1,527 posts
- 13. Alan Dershowitz 2,683 posts
- 14. Check Analyze 2,399 posts
- 15. Barcelona 133K posts
- 16. Token Signal 8,550 posts
- 17. Gonzaga 4,102 posts
- 18. Dialyn 7,544 posts
- 19. #FTTB 5,947 posts
- 20. Market Focus 4,659 posts
Something went wrong.
Something went wrong.