🔥Telegram İfşa

We are excited to release the dataset from the 2022 MABe Challenge! 🐭🪰 Our dataset consists of mouse (9 mil frames) and fly (4 mil frames) social interactions for studying behavioral representation learning! Paper: arxiv.org/pdf/2207.10553… Challenge: aicrowd.com/challenges/mul…

1/3) What is the best way to adapt large pre-trained vision models to downstream tasks in terms of effectiveness and efficiency? Drawing inspiration from the recent advances on Prompting in NLP, we propose a new simple and efficient method: Visual Prompt Tuning (VPT) 👇

I’m on the job market! I develop AI for scientists, to accelerate discovery from data & domain knowledge. My work tackles challenges from real-world workflows in domains such as neuroscience & healthcare, including annotation efficiency, interpretability & structure discovery.

Attending #ECCV2022? Drop by the poster session 1.A (poster 084) about VPT! 🔥❄️ w/ @CornellCIS, Belongie Lab and @MetaAI !

Best AI skillset in 2018: PhD + long publication record in a specific area Best AI skillset in 2023: strong engineering abilities + adapting quickly to new directions without sunk cost fallacy Correct me if this is over-generalized, but this is what it seems like to me lately

~1 month left to submit a paper to our workshop on Multi-Agent Behavior #CVPR2023! Come discuss multi-agent behavior, including biological and artificial agents, across wide ranges of spatial and temporal scales 🔬🐭🚶🪰🚗🏀🌍 Hope to see you in June!

Bard is now available in the US and UK, w/more countries to come. It’s great to see early Google AI work reflected in it—advances in sequence learning, large neural nets, Transformers, responsible AI techniques, dialog systems & more. You can try it at bard.google.com

A bunch of people have requested the slides for my "Scholars & Big Models" CVPR workshop talk. I didn't have a script, but I wrote a rough version of what I probably said at the bottom of each slide. Feedback is welcome! jonbarron.info/data/cvpr2023_…

VideoGLUE: Video General Understanding Evaluation of Foundation Models paper page: huggingface.co/papers/2307.03… We evaluate existing foundation models video understanding capabilities using a carefully designed experiment protocol consisting of three hallmark tasks (action

📢 Our #SMART101 challenge is now open! 🎉 Join the brightest minds in multimodal reasoning and cognitive models of intelligence to drive AI progress. 🚀 Don't miss out! Challenge closes on Sept. 1. Winning teams will receive prizes! 🏆 eval.ai/web/challenges… #VLAR #ICCV2023 #AI

Google presents Video Instruction Tuning Distilling Vision-Language Models on Millions of Videos paper page: huggingface.co/papers/2401.06… Experiments show that a video-language dual-encoder model contrastively trained on these auto-generated captions is 3.8% better than the

Introducing VideoPrism, a single model for general-purpose video understanding that can handle a wide range of tasks, including classification, localization, retrieval, captioning and question answering. Learn how it works at goo.gle/49ltEXW

Introducing Long Zhao, a Senior Research Scientist at Google, who worked to build VideoPrism: A Foundational Visual Encoder for Video Understanding. Read the blog to explore innovations in video understanding tasks and more →goo.gle/44vfn9D

Happy to share our recent work "Epsilon-VAE", an effective autoencoder that turns single-step decoding into a multi-step probabilistic process. Please check our paper for more detailed results! arXiv page: arxiv.org/abs/2410.04081

Introducing our latest work Video Creation by Demonstration, a novel video creation experience. Paper: arxiv.org/abs/2412.09551 Project: delta-diffusion.github.io Huggingface: huggingface.co/papers/2412.09…

After over 15 months, we are excited to finally release VideoPrism! The model comes in two sizes, Base and Large, and the video encoders are available today at github.com/google-deepmin…. We are also working towards adding more support into the repository, please stay tuned.

Excited! VideoPrism-Base/Large are publicly available now: github.com/google-deepmin… Check it out if you need a versatile video encoder for video-language or video-native tasks. Feedback appreciated!

At 4:00 today, stop by the #CVPR2025 Google booth where Ting Liu will demo a model for video creation by demonstration that can generate physically plausible video that continues naturally given a context scene. Find sample videos at delta-diffusion.github.io

Excited to share the release of VideoPrism! 🎥 📏Generate video embeddings 👀Useful for classifiers, video retrieval, and localization 🔧Adaptable for your tasks Model: hf.co/google/videopr… Paper: arxiv.org/abs/2402.13217 GitHub: github.com/google-deepmin…

And great to see VideoPrism makes the 1000th!