🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Presenting the CONTAINER (CONtext Aggregation NEtwoRk) -- a unified view of Transformers, CNNs, MLP-Mixers via a general-purpose building block for multi-head context aggregation. arxiv.org/abs/2106.01401 by Peng, Jiasen Lu, Hongsheng, Roozbeh Mottaghi, Ani Kembhavi Ai2

So excited for this, can't wait to see how our AI does on the strategic gem Polytopia, one of my favourite mobile games!!

1/ Really excited to share: Language in Brains, Minds, and Machines w Nancy Kanwisher @[email protected] Ev (like in 'evidence', not Eve) Fedorenko 🇺🇦 Annual Reviews We survey the insights that language models (LMs) provide on the question of how language is represented and processed in the human brain. rb.gy/8afztv

A pleasure chatting with Mitch Waldrop about brains and LLMs, including the use of LLMs as "electronic lab rats" (and much more, featuring awesome work from various labs in the recent years)

Introducing Poliformer: A large transformer agent trained with on-policy RL that achieves 85% Object Nav success across two embodiments in Sim and Real. Does this plot remind you of anything ? It reminds me of the DDPPO plot that solved PointNav. Now after 4 years, we finally

Our new paper "Open Problems in Technical AI Governance" led by Ben Bucknall & me is out! We outline 89 open technical issues in AI governance, plus resources and 100+ research questions that technical experts can tackle to help AI governance efforts🧵 t.ly/Y-mQ1

Excited to present our #ECCV2024 paper on “Understanding Physical Dynamics with Counterfactual World Modeling” with Honglin Chen, Kevin Feigelis, Daniel Bear and Daniel Yamins. Come by poster 260 at the exhibition area from 430-630pm today European Conference on Computer Vision #ECCV2026 TLDR: We introduce Counterfactual

Attending #ECCV2024 in Milan? Stop by poster 260 to see our recent work on understanding physical dynamics with world modeling! Unfortunately, I cannot be there due to visa issues, but my amazing collaborator Rahul Venkatesh will be there - stop by and chat with him!

Accurate and controllable scene generation has been difficult with natural language alone. You instead need a language for scenes. Introducing the Scene Language — a visual representation for high-quality 3D/4D generation by integrating programs, words, and embeddings — 🧵(1/6)

📢Excited to share: I'm again leading the efforts for the Responsible AI chapter for Stanford University's 2025 AI Index, curated by Stanford HAI. As last year, we're asking you to submit your favorite papers on the topic for consideration (including your own!) 🧵 1/

Modern generative models of images and videos rely on tokenizers. Can we build a state-of-the-art discrete image tokenizer with a diffusion autoencoder? Yes! I’m excited to share FlowMo, with Kyle Hsu, Justin Johnson, Fei-Fei Li, Jiajun Wu. A thread 🧵:

We just finished up Winter quarter CS375: Large-Scale Neural Network Models for Neuroscience. Check out the publicly available Syllabus and lecture notes cs375.stanford.edu/course-calenda…

New paper on self-supervised optical flow and occlusion estimation from video foundation models. Stefan Stojanov Jiajun Wu Seungwoo (Simon) Kim Rahul Venkatesh tinyurl.com/dpa3auzd @

This week's #PaperILike is "Impossibly Good Experts and How to Follow Them" (Walsman et al., ICLR 2023). A very well written paper that will be especially interesting if you're using teacher-student training in your work. PDF: openreview.net/pdf?id=sciA_xg…

A key finding is that strong patch locality constraints force the model to learn how scenes look from the patch level up, making it less prone to generating global artifacts in novel view synthesis, or changing object identity – which are known problems with diffusion models.

After 7 years, thrilled to finally share our #MICrONS functional connectomics results! We recorded activity from ~75K neurons in visual cortex in a single mouse, then mapped its wiring using electron microscopy. To systematically characterize neuron function, we built the first

Exciting news! Open Problems in Technical AI Governance was published at TMLR! This is a good time to mention that we're organizing a workshop on TAIG at #ICML2025. Papers are due May 7th and we're also looking for PC members Check out Technical AI Governance @ ICML 2025 for more details!