Brandon Amos

Excited to release AlgoTune!! It's a benchmark and coding agent for optimizing the runtime of numerical code 🚀 algotune.io 📚 algotune.io/paper.pdf 🤖 github.com/oripress/AlgoT… with @OfirPress @ori_press @PatrickKidger @b_stellato @ArmanZharmagam1 & many others 🧵

[1/9] We created a performant Lipschitz transformer by spectrally regulating the weights—without using activation stability tricks: no layer norm, QK norm, or logit softcapping. We think this may address a “root cause” of unstable training.

optimization theorem: "assume a lipschitz constant L..." the lipschitz constant:

I'm excited to share new optimality conditions for nonlinear stochastic optimal control, and the first indirect shooting method for solving these problems! 📖 arxiv.org/abs/2502.06726 💡 How? Using rough path theory ⬇️

Here's how my recent papers & reviews are going: * To solve a vision problem today, the sensible thing is to leverage a pre-trained VLM or video diffusion model. Such models implicitly represent a tremendous amount about the visual world that we can exploit. * Figure out how to…

🚨 The era of infinite internet data is ending, So we ask: 👉 What’s the right generative modelling objective when data—not compute—is the bottleneck? TL;DR: ▶️Compute-constrained? Train Autoregressive models ▶️Data-constrained? Train Diffusion models Get ready for 🤿 1/n

1/N I’m excited to share that our latest @OpenAI experimental reasoning LLM has achieved a longstanding grand challenge in AI: gold medal-level performance on the world’s most prestigious math competition—the International Math Olympiad (IMO).

📢Presenting SDE Matching🔥🔥🔥 🚀We extend diffusion models to construct a simulation-free framework for training Latent SDEs. It enables sampling from the exact posterior process marginals without any numerical simulations. 📜: arxiv.org/abs/2502.02472 🧵1/8

Autoregressive models are too restrictive by forcing a fixed generation order, while masked diffusion is wasteful as it fits all possible orders. Can our model dynamically decide the next position to generate based on context? Learn more in our ICML paper arxiv.org/abs/2503.05979

Soaking up the sun (surprisingly) at #ICML2025. Come check out our work on Generative Modeling on the space of distribution, with cool applications to spatial genomics. With @brandondamos, Aram Pooladian, & @dana_peer. 🗓️ Today, 4:30-7pm 📍 East Exhibition Hall, #3301

For everyone interested in precise 📷camera control 📷 in transformers [e.g., video / world model etc] Stop settling for Plücker raymaps -- use camera-aware relative PE in your attention layers, like RoPE (for LLMs) but for cameras! Paper & code: liruilong.cn/prope/

Thrilled to share the Community Alignment dataset -- the product of a massive collaborative effort with so many awesome folks. Can't wait to see the future research it unlocks!

As one of the people who popularized the field of diffusion models, I am excited to share something that might be the “beginning of the end” of it. IMM has a single stable training stage, a single objective, and a single network — all are what make diffusion so popular today.

Tokenization has been the final barrier to truly end-to-end language models. We developed the H-Net: a hierarchical network that replaces tokenization with a dynamic chunking process directly inside the model, automatically discovering and operating over meaningful units of data

🚀 Introducing PhysiX: One of the first large-scale foundation models for physics simulations! PhysiX is a 4.5B parameter model that unifies a wide range of physical systems, from fluid dynamics to reaction-diffusion, outperforming specialized, state-of-the-art models.

Personalization methods for LLMs often rely on extensive user history. We introduce Curiosity-driven User-modeling Reward as Intrinsic Objective (CURIO) to encourage actively learning about the user within multi-turn dialogs. 📜 arxiv.org/abs/2504.03206 🌎 sites.google.com/cs.washington.…