ML Research Engineer @Apple | interpretability, alignment, compositionality of diffusion models | EX - @AdobeFirefly, @SonyAI_global
Santa Clara, CA
Joined August 2021
- Tweets74
- Following423
- Followers147
- Likes406
Pinned Tweet
Most diffusion research today asks: How can we sample faster?
But I think another question is equally important: Are we training diffusion models in the right way?
famous-bubbler-dcc.notion.si…
Xianghao Kong retweeted
A talk based on my recent ICML paper 'On the Separability of Information in Diffusion Models' is now available on YouTube: youtube.com/watch?v=8kGh6W-9…
(1/n)
That’s terrific! Can’t waiting to reading it!!
Our book "Generative AI and Stochastic Thermodynamics: A Tale of Free Energies" is out this month. Over the coming weeks @sirui_lu97, @wellingmax and I will post about some of the topics it covers.
Starting it off: heat and work in latent variable models and variational EM.
This may matter even more for recurrent diffusion and interactive generative models, where small errors can accumulate over long-term rollouts.
Xianghao Kong retweeted
Today we're releasing ZAYA1-74B-Preview, a major milestone in scaling pretraining on @AMD.
ZAYA1-74B-Preview is a 4B active / 74B total MoE.
This preview model is a strong pre-RL base checkpoint. The final post-trained reasoning model is coming soon. 🧵
Xianghao Kong retweeted
Really excited to see Uni-1 out in the world 🔥Our first unified model.
The range of things this model can do is wild: image-to-~100 styles, manga generation, multi-ref with strong identity preservation, temporal storytelling, sketch-to-image, spatial reasoning, multilingual infographics, layering… the capability range is honestly unreal. this is just the start 🫡 check out the blog to learn more lumalabs.ai/uni-1
Proud of the team and what we’re building at @LumaLabsAI 🚀
Introducing Uni-1, Luma’s first unified understanding and generation model, our next step on the path towards unified general intelligence.
lumalabs.ai/uni-1
ALT Combine the black and white curly-haired dog with pink bandana, the Boston Terrier in plaid harness, and the black-and-white cat into a single scene where they are dressed in academic regalia, standing before a whiteboard filled with scientific diagrams and text, with the Luma AI logo placed in the top-left corner.
I’m currently in transit to San Diego for NeurIPS. If you’re also killing time, feel free to check out a 2-minute-30-second horror sci-fi short film Michael and I recently created. We’d love any comments or likes:
devpost.com/software/dreamca…
Looking forward to catching up at the venue! 🎥
I feel the debate shouldn’t only be about whether DiT is effective, but also about how information preservation is the key to accelerating diffusion training. Our MicroDiT (arxiv.org/abs/2407.15811) paper showed this: by letting masked token info mix into unmasked ones, we can cut down a lot of tokens with only minor performance loss.
Interestingly, two months ago, when I caught up with @StefanABaumann at #CVPR, we discussed how TREAD and MicroDiT are conceptually similar from info perspective. Maybe it’s time to look at diffusion through an information-theoretic lens: from post-training (for the better alignment) to latent space curation, I believe this could lead to some really exciting discoveries!
Xianghao Kong retweeted
Introducing our V1 Video Model. It's fun, easy, and beautiful. Available at 10$/month, it's the first video model for *everyone* and it's available now.
Xianghao Kong retweeted
+ @cveu_workshop starting at 1:00 PM, 207 A-D.
Very happy to be in Music City for #CVPR2025 My lab is presenting 7 papers, 4 selected as highlights. My amazing students @IrohXu @zixuan_huang @Wenqi_Jia @bryanislucky Xiang Li @fionakryan and postdoc Sangmin Lee are here! @siebelschool @uofigrainger
Heading to Nashville 🎸 for @CVPR (06/11 - 06/16)!
Always excited to catch up with old friends and make new connections. Let’s grab a coffee ☕️ or chat about diffusion models, post-training, or just life!
#CVPR2025 #Diffusion #GenerativeAI #Nashville
