Check out the new addition to our TarFlow franchise. TLDR: normalizing flows “just work” for generating videos. This adds another strong evidence to our argument that NFs are capable generative models; and I’m now more convinced than ever that they will continue working better.

I’ll be at San Diego attending #NeurIPS2025 Dec 3-7. DM me if interested in diffusion model, multimodal, protein generative models! We’re looking for FTE to join us working on generative models. You can also find me at Apple  booth on Dec 3 3-5pm.

At #NeurIPS2025 Tue-Fri presenting 3 papers from our 🍎Apple ML research team. Interested in LLM, RL, reasoning, and diffusion LLMs. We also have FY26 research intern and full-time positions available. DM me if interested for a chat!

Thanks @_akhaliq for sharing our work!! All the code for STARFlow-V and the prior work STARFlow (NeurIPS spotlight @ this Thur) has been released https://t.co/ATU9XtaKio We also have the T2I model weights available now https://t.co/DSeFXRYAqZ Let's push more on Scalable NFs!

Checkout our new work!

LLMs are notorious for "hallucinating": producing confident-sounding answers that are entirely wrong. But with the right definitions, we can extract a semantic notion of "confidence" from LLMs, and this confidence turns out to be calibrated out-of-the-box in many settings (!)

Might also be interested in checking our TARFlow series! TARFlow: https://t.co/Gb7NETqEw2 ICML2025 Oral STARFlow: https://t.co/bpkY7SYx4z NeurIPS2025 Spotlight TARFlow-LM: https://t.co/BLHoXt9m5Q NeurIPS 2025 … and more maybe soon🤖

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

We’re excited to share our new paper: Continuously-Augmented Discrete Diffusion (CADD) — a simple yet effective way to bridge discrete and continuous diffusion models on discrete data, such as language modeling. [1/n] Paper: https://t.co/fQ8qxx4Pge

Lastly, great thanks for all of my amazing collaborators! @D_Berthelot_ML @zhaisf @jmsusskind @thoma_gu @UnderGroundJeg

Thus, we are able to achieve better performance with same limited sampling budget :) For more experiments results please refer to the paper.

The generated results remain diverse even when the initial noise fed into the network is identical and the dynamics are solved by an ODE solver under the same SNR discretization.

The difference arises in sampling: the momentum system is a pseudo-SDE (loosely speaking), where dynamics are deterministic but stochasticity emerges from variable interactions. Compared with the discretized Wiener process (left), such pseudo noise is smoother.

First of all, they are identical to vanilla diffusion models during training, even with additional input variables. It means, 1. There is no training benefit for augmented system. 2. A Pre-trained diffusion model can even be DIRECTLY used for sampling in these augmented systems.

Here is our niche accepted paper ( https://t.co/dJoM1UitWk) in NeurIPS2025 :) This is also a summary of previous augmented diffusion models (PFGM, CLD, AGM, etc.), and investigates in which cases, if any, they are actually useful in practice.

Three generative modeling papers from my team accepted to #NeurIPS2025, two on TarFlow and one on Diffusion. 1. StarFlow (Spotlight) https://t.co/ynB2z4Xgfa, scales TarFlow in latent space and demonstrates unprecedented sample quality from pure NF models. Work led by @thoma_gu.

Excited to STARFlow has been accepted at #NeurIPS2025 as a **Spotlight** paper! Super excited and looking forward to seeing more research directions on scalable normalizing flows as an alternative to this existing diffusion world!🧐 Huge congrats to my amazing collaborators!!

Normalizing Flows are coming back to life. I'll be attending #ICML2025 on Jul 17 to present TarFlow -- with the Oral presentation at 4:00 PM in West Ballroom A and the poster at 4:30 PM in East Exhibition Hall A-B #E-2911. Stop by and I promise it will be worth your time.

@shaulneta @urielsinger @itai_gat @lipmanya Congrats on Amazing work! Do you think it may be interesting to mention some possible connections to our previous work DART https://t.co/cNf6RjaWbL 👀

Normalizing flows have some inherent limitations. However, if some of these challenges can be mitigated, this direction may still hold promise as a competitive and scalable generative modeling. Here is our another attempt in this direction.