Victor Fragoso 💻

We are happy to announce the release of the Phi 3.5 family of models! Check them out: 👉Phi 3.5 mini: huggingface.co/microsoft/Phi-… 👉MoE: huggingface.co/microsoft/Phi-… 👉Phi 3.5 vision: huggingface.co/microsoft/Phi-…

🚀MIT Flow Matching and Diffusion Lecture 2026 Released (diffusion.csail.mit.edu)! We just released our new MIT 2026 course on flow matching and diffusion models! We teach the full stack of modern AI image, video, protein generators - theory and practice. We include: 📺 Videos: Step-by-step derivations. 📝 Notes: Mathematically self-contained lecture notes 💻 Coding: Hands-on exercises for every component We fully improved last years’ iteration and added new topics: latent spaces, diffusion transformers, building language models with discrete diffusion models. Everything is available here: diffusion.csail.mit.edu A huge thanks to Tommi Jaakkola for his support in making this class possible and Ashay Athalye (MIT SOUL) for the incredible production! Was fun to do this with @RShprints! #MachineLearning #GenerativeAI #MIT #DiffusionModels #AI

🚀 Today we’re releasing FlashOptim: better implementations of Adam, SGD, etc, that compute the same updates but save tons of memory. You can use it right now via `pip install flashoptim`. 🚀 arxiv.org/abs/2602.23349 A bunch of cool ideas make this possible: [1/n]

#CVPR2026 final decisions are out! Available for now only via email. Good luck🤞

Reminder: the #ECCV2026 submission portal is now open.

The day has arrived, somewhere in the world 😅 Our Program Chairs and senior Area Chairs are going through the final checks. Only the Program Chairs know the exact timing of the release. The rest of us will find out at the same time. Good luck!

Planning to come to #CVPR2026 Denver? Check out our official conference hotel options. cvpr.thecvf.com/Conferences/20…

Tu camino para estudiar en EE. UU. comienza aquí. 🫵🏼 La Community College Fair llega muy pronto ✨ 📅 Martes, 17 de marzo de 2026 🎓 Guarda la fecha y mantente atento a más información. #EducationUSAMx

Call for Papers: eccv.ecva.net/Conferences/20…

If there are major concerns raised in the reviews for your #CVPR2026 submission and you don’t plan to submit a rebuttal, then withdraw the paper now.

Please double-check the #CVPR2026 rebuttal deadline. The deadline is FIRM! NO extensions and NO exceptions. Good luck! Dates: cvpr.thecvf.com/Conferences/20…

Que nada te detenga. 🚀 La información abre puertas. 🔑 EducationUSA México 🇺🇸✨ 🔗 bit.ly/centros_mex #EducationUSAMx

#CVPR2026 reviews are slowly being dispatched over email. Good luck!

We are hiring for a researcher in the foundations of generative AI: apply.careers.microsoft.com/careers/job/19… . Please share with whoever may be interested.

Following our previous #ICML2026 call for tutorial topics, we now welcome tutorial proposals! 📆Deadline: February 6, but proposals will be reviewed as they arrive Courtesy the tutorial chairs Claire Vernade (@vernadec) and Adam White

Nominations for the organizing committee for NeurIPS 2026 are now open! blog.neurips.cc/2026/01/07/neu…

The reviewing phase closes Jan 12, 2026. Please submit all feedback via the “Official Review” button (not “Official Comment”). Reviewers with submissions who miss their assigned reviews may risk desk rejection. Thank you for helping keep #CVPR2026’s review process timely.

CVPR 2026 Registration is now live! Please remember if you have a paper accepted, you must register as an author. Anyone is welcome to attend CVPR as long as they are registered before arriving onsite. cvent.me/dMoWob

#CVPR2026 submissions have now been assigned to reviewers. Regular reviewers received at most four papers. Please check your email for next steps.

Video diffusion models have strong implicit representations of 3D shape, material, and lighting, but controlling them with language is cumbersome, and control is critical for artists and animators. GenLit connects these implicit representations with a continuous 5D control signal describing the direction and intensity of a point light source. This enables single-image near-field relighting of an image using a video diffusion model. We use a ControlNet-like approach and show that, with a small amount of synthetic data, GenLit generalizes to complex real-world images. Given a single image and the 5D lighting signal, GenLit creates a video of a moving light source that is inside the scene. It moves around and behind scene objects, producing effects such as shading, cast shadows, secularities, and interreflections with a realism that is hard to obtain with traditional inverse rendering methods. GenLit shows that it is possible to get continuous control over implicit physical processes within a video model. I think this is just the beginning and promises to make such models much more practical for creators. @sbharadwaj__ will present today at SIGGRAPH Asia Room: S423/S424, Level 4 @ 13:50 on 15 of Dec. genlit.is.tue.mpg.de

Visit booth #1703 during #NeurIPS in San Diego to engage with Microsoft experts, meet our recruiters, and check out our innovations and demos. Explore the schedule here: msft.it/6013tkmdX #MSFTatNeurIPS