Stanford AI Lab (@StanfordAILab) unfollowed @percyliang on Oct 12, 2024

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…

RT @irena_gao: Applying to Stanford's CS PhD program? Current graduate students are running a SoP + CV feedback program for URM applicants…

RT @MayeeChen: honored to have contributed to this incredibly exciting work! Archon shows how intelligently blending concepts like ensembli…

RT @sanmikoyejo: New website!

@honglin_c @recursus @dyamins @eccvconf I am very grateful to work on this project with amazing collaborators from Stanford @NeuroAILab , MIT and UCSD – @honglin_c , Kevin Feigelis, @recursus, @KlemenKotar, @KJedoui, Wanhee Lee, Felix Binder, Sherry Liu, @realkevinsmith, @judyefan and @dyamins (5/5)

@honglin_c @recursus @dyamins @eccvconf These structures help achieve state-of-the-art results the Physion benchmark for physical dynamics understanding (4/5) https://t.co/DkSRITYCjy

RT @dyamins: Really excited about this line of work! If you're at ECCV, go see Rahul's poster.

RT @Rahul_Venkatesh: Excited to present our #ECCV2024 paper on “Understanding Physical Dynamics with Counterfactual World Modeling” with @h…

RT @Qi_Wu577: Introducing Helpful DoggyBot🐕, a legged mobile manipulation system: - A quadruped with a mouth - Agile whole-body skills like…

RT @zipengfu: Introduce DoggyBot🐕series: quadrupeds can also do manipulation. It's a fruitful 4yr journey working on robot dogs from walki…

RT @JonSaadFalcon: What is the best way to spend your inference compute budget to create LLM systems greater than the sum of their parts?…

RT @leg2015: One ocean, two planes, three trains, and 35 stickers later (long story), my first ECCV poster is finally on the books! https:/…

RT @KarlPertsch: I'll give a talk at the Multimodal Agents workshop at ECCV tomorrow Sept 30, at 2:20pm CET. Excited for my first talk at a…

RT @leto__jean: This is a great paper with actionable suggestions on how to evaluate robot policies. If everyone were to follow these, we…

RT @sunfanyun: Training RL/robot policies requires extensive experience in the target environment, which is often difficult to obtain. How…

RT @SuryaGanguli: Our new NeurIPS spotlight: https://t.co/0GcC4xz57K Get rich quick: new exact solutions to learning dynamics reveal how un…

RT @StanfordHAI: .@StanfordHAI Co-Director Fei-Fei Li joined @BillGates, Francis Kere, and @BillClinton for the main stage closing session…

RT @Dilip_Arumugam: A belated life update: I defended my PhD and graduated from @StanfordAILab! Immensely grateful to my advisor, BVR, and…

RT @deedydas: Stanford Computer Science has so much alpha because its classes update within months of real-world research. Most other top…