Stanford AI Lab

RT @RylanSchaeffer: Excited to announce our Vision-Language Models (VLM) image jailbreak paper will be a spotlight at #neurips2024 workshop…

RT @IanHuang3D: Andy Warhol. Michelangelo. Rembrandt. They all had assistants. Let AI be your #Blender technician so that you can do more…

project website: https://t.co/tKQR8hUFmM Big thanks to all my amazing collaborators: @hariniskumar10, @yiangela7, Yihan Zhou, @zookae, Jonathan Tremblay, @locross, @jiajunwu_cs, @nickhaber! 7/7

RT @vamsi_nimmala: Time has proven this tweet right. @chrmanning said it best—excellent at NLP, but still far from true understanding.…

RT @YangjunR: When we released the observational scaling law paper (https://t.co/GrfiKD9fQc), we preregistered our predictions of future mo…

RT @RogerDai1217: Why hand-engineer digital twins when digital cousins are free? Check out ACDC: Automated Creation of Digital Cousins 👭 f…

RT @agiachris: In real-world deployment settings, generative robot policies are bound to encounter out-of-distribution scenarios that cause…

RT @drmapavone: Check out our latest work on monitoring data-driven robot policies: Sentinel - a runtime monitor that detects unknown failu…

RT @Avanika15: Beyond honored to have ECLAIR featured in @nathanbenaich’s amazing 2024 @stateofaireport As always, the report is amazing —…

RT @Allen_A_N: LLMs are in-context RL learners, but not great because they can’t explore well. How do we teach LLMs to explore better? 🤔 🔮…

RT @RogerDai1217: Why hand-engineer digital twins when digital cousins are free? Check out ACDC: Automated Creation of Digital Cousins 👭 f…

RT @stephentian_: Learned visuomotor robot policies are sensitive to observation viewpoint shifts, which happen all the time. Can visual pr…

RT @tatsu_hashimoto: I've been excited about pre-registration as a way to do better science for model scaling. Observational scaling holds…

We are also happy to share that the paper has been accepted at #NeurIPS2024 as a spotlight🔥 For more details about the paper, see the initial thread👇 https://t.co/gS9yBLOAS6

RT @RogerDai1217: Why hand-engineer digital twins when digital cousins are free? Check out ACDC: Automated Creation of Digital Cousins 👭 f…

RT @Koven_Yu: 🔥Spatial intelligence needs fast, *interactive* 3D world generation 🎮 — introducing WonderWorld: generating 3D scenes interac…

RT @RogerDai1217: Why hand-engineer digital twins when digital cousins are free? Check out ACDC: Automated Creation of Digital Cousins 👭 f…

RT @1x_tech: Introducing NEO Beta. Designed for humans. Built for the home. https://t.co/5S6jpRjUQp

RT @RogerDai1217: Why hand-engineer digital twins when digital cousins are free? Check out ACDC: Automated Creation of Digital Cousins 👭 f…

RT @emilyzjin: Everyday, we solve a number of "whodunit" mysteries: Who left the fridge open? Who spilled the food? Who turned on the light…

How to govern open(-weight) foundation models has been a hot topic over the last year. This @ScienceMagazine article captures our thinking on benefits (let's not forget about this!) and risks of release model weights. We promoted the "marginal risk" framing that is now fairly

A new leaderboard for VLMs built using the HELM framework, featuring 21 datasets (including MMMU, RealWorldQA, VQAv2, GQA, Bingo, etc.), evaluating your favorite VLMs using standard prompts. All predictions available on our website for full transparency!

RT @SISLaboratory: Many problems, such as rover exploration, can be framed as informative path planning. @joshua_0tt tackles the problem of…

I totally agree! This study is just the start and I'm looking forward to the future. I talk about the exciting challenges and pressing need for multi-modal interaction capabilities in our Limitations section (screenshot for reference)! https://t.co/G2Orw7x1a9

RT @monicaglee: So excited about this study finally being out in the world 🎉 and grateful to have the wonderful @rose_e_wang on the @NSSAcc…

RT @adn_twitts: Excellent talk by @sanmikoyejo What a difference to have a technical talk and not an ad 🎉 Excellent selection on this on…

RT @ducha_aiki: Keynote on distribution shift by @sanmikoyejo #ECCV2024 https://t.co/rt2WMJOeOh

RT @SuryaGanguli: If you are interested in learning the “dark calculus of the New Rationalists” - I teach a course on it 🤣 https://t.co/XLe…

RT @StanfordOnline: Join us for an exclusive fireside chat with Stanford experts @AndrewYNg and @mehran_sahami , where they will provide v…

RT @Zhang_Yu_hui: Auto-regressive text-to-image generation using VQGAN shows promise in AI art. 🎨 Our #EMNLP2024 study reveals: pre-traine…

RT @larsbrinkhoff: #OnThisDay 33 years ago, the @StanfordAILab KL10 was shut down and dismantled. https://t.co/9RdFSI7DqK

RT @7etsuo: 🧵1/2 Learn about deep generative models like VAEs, GANs, autoregressive models, and more, and how they’re applied in computer v…

RT @AdamWHarley: Come to poster 263! Haodi and I are presenting "Ultrametric Feature Fields", combining SAM and NeRFs to get clean 3D hiera…