Stochastic and deterministic sampling strategies for diffusion models produce... strikingly different trajectories, but both ultimately achieve the same aim. I had a great time presenting our work, Diffusion Explorer, this week at IEEE VIS in Vienna.show more

Diffusion models leverage a variety of samplers. Deterministic methods... like DDIM produce orderly paths. In contrast, stochastic samplers like DDPM produce chaotic trajectories. Despite their differences, both methods draw valid samples from the underlying distribution.show more

“For the first time in a long time, I... don’t see nothing but Americans on this work site. We making America great again, one work site at a time.” Love this!! 🔥👇🏼show more

This is the coolest Diffusion work I've seen in... a while! It generates Visual Anagrams, a type of optical illusion where an image looks like one thing, but changes appearance when transformed. It works with any orthogonal transformation matrices, which luckily include rotation, permutation (jigsaw puzzles), and color negation. Intuitively, the method first inverts the noise from multiple image transforms (with different text prompts), and then average them. After taking a diffusion step in the averaged noise, the resulting image becomes an anagram that aligns with the texts in different views. It does very little computation, using pre-trained Stable Diffusion. Simple, elegant, and inexpensive technique for non-professionals to create some interesting art! Paper: Website: It's open-source! Authors: Daniel Geng, Inbum Park, Andrew Owens.show more

I had a great time at FN this morning... in Columbia. I will always stand strong for the Second Amendment.show more

World modeling and imitation learning have largely been considered... two disparate worlds. In our recent work, Unified World Models, just accepted to #RSS2025, Chuning Zhu provides a dead-simple unifying solution: just train a joint diffusion model over actions and future states, but with *decoupled* diffusion time steps across these modalities. Manipulating these decoupled time steps then allows for marginalization or conditioning on actions or states; a single model can serve as a policy, forward dynamics model, video prediction model, or inverse dynamics model by simply setting diffusion timesteps carefully. The resulting model can leverage video datasets along with robot training data much more effectively, and shows improved robustness, generalization, and flexibility. This is exciting because it is frustratingly simple, scalable, and shows strong improvement on real-world robotics problems. Please refer to Chuning Zhu 's excellent thread for more details! More details/code can be found on our website and in the paper -show more

got in trouble at work for being on X... too much, so I built a tool that scrolls my timeline the more I code. now I can do both at the same timeshow more

watching this just makes me tear up a bit... it truly amazes me how got7 manages their time juggling both group and solo schedules… they can be spotted at the airport together, same day same time but for different flights it’s so bittersweetshow more

If you’re not willing to work weekends, late nights,... and sacrifice your free time for a few years.. Then don’t be surprised when you’re in the same place five years from now. Everyone wants the results, but nobody wants to put in the work when it’s inconvenient. You can’t have both comfort and success at the same time.show more

We release Diamond Maps💎 unlocking accurate and efficient guidance... for diffusion models. Our experiments show that our methods scale incredibly well. Excited to see what people will build with this! Accurate guidance has been a notoriously hard problem, but in this work, we’re bringing TWO (!) solutions to the table. The recipe for success: 1️⃣ Speed: Use distilled models (flow maps, mean flows, consistency models). 2️⃣ Exploration: Inject stochasticity to properly explore your search space. Because this fundamentally improves anything using flow matching and diffusion, we see a lot of potential for applications across audio, robotics, molecules, and beyond. Paper: Code: Huge thanks to an amazing team: Douglas Chen, Luca Eyring, Ishin Shah, Giri Anantharaman, Yutong (Kelly) He, Zeynep Akata, Tommi Jaakkola, Nicholas Boffi, and Max Simchowitz. It was awesome bringing this to life together!show more

We finished our week in Toronto with a 2-3... record. Not enough for playoffs, but we had a great time at the GSOC Players’ Championship! Thank you icemakers, sponsors, fans, and volunteers for hosting a fantastic event. A big thank you to 𝓀𝓇𝓎𝓈𝓉𝑒𝓃 for joining us!show more

junseo to leo 🩷 🦊 if i had to... describe our leo in just one keyword he’s a music genius and a reliable leader 🦊 of course he’s also great at dancing 🦊 but honestly i think it’s really cool how good he is at both rapping and singing 🦊 and since he also does such a great job leading our team that’s why i chose this keyword for himshow more

Wonderland: Navigating 3D Scenes from a Single Image Contributions:... • First, we introduce a representation for controllable 3D generation by leveraging the generative priors from camera-guided video diffusion models. Unlike image models, video diffusion models are trained on extensive video datasets. This enables them to capture comprehensive spatial relationships within scenes across multiple views and embed a form of "3D awareness" in their latent space, which allows us to maintain 3D consistency in novel view synthesis. • Second, to achieve controllable novel view generation, we empower video models with precise control over specified camera motions. We introduce a novel dual-branch conditioning mechanism that effectively incorporates desired diverse camera trajectories into the video diffusion model. This enables expansion of a single image into a multi-view consistent capture of a 3D scene with precise pose control. • Third, to achieve efficient 3D reconstruction, we directly transform video latents into 3DGS. We propose a novel latent-based large reconstruction model (LaLRM) that lifts video latents to 3D in a feed-forward manner. With this design, during inference, our model directly predicts 3DGS from a single input image, effectively aligning the generation and reconstruction tasks—and bridging image space and 3D space—through the video latent space. Compared with reconstructing scenes from images, the video latent space offers a 256× spatial-temporal reduction while retaining essential and consistent 3D structural details. Such a high degree of compression is crucial, as it allows the LaLRM to handle a wider range of 3D scenes within the reconstruction framework, with the same memory constraints.show more

I’m a proud Czech, I’m a proud European. I’m... proud to be both at the same time and for this I will always be thankful.show more

I had an amazing time at Twitchcon and it... was great meeting everyone, thank you all for the good memories🫶 I didnt take pictures but I got this bangershow more

Create a 3D model from a single image, set... of images or a text prompt in < 1 minute 😮‍💨 This new AI paper called CAT3D shows us that it’ll keep getting easier to produce 3D models from 2D images — whether it’s a sparser real world 3D scan (a few photos instead of hundreds) or your favorite 2D image generator like Midjourney (just an image). How does this magic work? “This architecture is similar to video diffusion models, but with camera pose embeddings for each image instead of time embeddings. The generated views are passed into a robust 3D reconstruction pipeline to create the 3D representation (Zip-NeRF or 3DGS)”show more

The Hidden Language of Diffusion Models paper page: tackle... the challenge of understanding concept representations in text-to-image models by decomposing an input text prompt into a small set of interpretable elements. This is achieved by learning a pseudo-token that is a sparse weighted combination of tokens from the model's vocabulary, with the objective of reconstructing the images generated for the given concept. Applied over the state-of-the-art Stable Diffusion model, this decomposition reveals non-trivial and surprising structures in the representations of concepts. For example, we find that some concepts such as "a president" or "a composer" are dominated by specific instances (e.g., "Obama", "Biden") and their interpolations. Other concepts, such as "happiness" combine associated terms that can be concrete ("family", "laughter") or abstract ("friendship", "emotion"). In addition to peering into the inner workings of Stable Diffusion, our method also enables applications such as single-image decomposition to tokens, bias detection and mitigation, and semantic image manipulationshow more

🚀New paper out - We present Video-MSG (Multimodal Sketch... Guidance), a novel planning-based training-free guidance method for T2V models, improving control of spatial layout and object trajectories. 🔧 Key idea: • Generate a Video Sketch — a spatio-temporal plan with background, foreground, and motion in the pixel space. • Encode this structure directly into the latent space of the diffusion model during generation, which does not require fine-tuning or additional memory during inference. 🧵show more

Airport tram ride at a different airport on a... different day:) Traveling for work this week and I had an entire tram car to myself at San Francisco Airport this morning so I took a quick vid to show what it looks like inside. Las Vegas one looks pretty similar too. Not quite an attraction ride but I'm glad Seokjin enjoyed it! #Jin #방탄소년단진 #진 방탄소년단show more

Most recent diffusion language model research (that I’ve seen)... seems to be using masking as the noising process. It looks like, however, most closed-source models (Google Gemini Diffusion and possibly Inception Labs’ Mercury) use a different noising process, where instead of masking tokens, they replace them with different tokens (either with a random token or a semantically similar token). I wondered how they were getting such high throughput with the latter noising process, since I believed that optimizing inference with KVCache approximation would be more difficult (for various reasons). I visualized this noising process with tiny-diffusion and compared it to normal unmasking, and was very surprised to see how fast the generation “settles” into a reasonable output, and then only slightly refines afterwards, requiring much fewer steps in total. Unmasking (where tokens are never remasked, the typical implementation) is inherently limited in generation speed by the fact that an increase in tokens decoded per step leads to more errors due to the mismatch between individual and marginal token probability distributions we sample from. The token replacement noising process seems to have a much different set of characteristics. Because we sample each token per step, every token makes “progress” towards the final output each iteration (in addition to *potentially* giving other tokens more information in future steps). Generally, masking has outperformed other noising processes, which is probably why most research focused on it (using smaller models). But the paper referred to in the retweet shows that random replacement as a noising process may scale better as model size increases. Big labs might have noticed these results much earlier (due to having drastically more training resources and being able to test larger models), which may explain the discrepancy in the choice of noising process. I’m gonna test this with larger models, since tiny-diffusion only has 10M parameters.show more