Ting Liu
@_tingliu
Followers
243
Following
87
Media
1
Statuses
48
Researcher @GoogleDeepMind
Los Angeles, CA
Joined January 2010
And great to see VideoPrism makes the 1000th!
Check out the 999 open models that Google has released on @huggingface: https://t.co/Fo4Ycn9ARi (Comparative numbers: 387 for Microsoft, 33 for OpenAI, 0 for Anthropic).
0
0
4
Excited to share the release of VideoPrism! 🎥 📏Generate video embeddings 👀Useful for classifiers, video retrieval, and localization 🔧Adaptable for your tasks Model: https://t.co/B2RPZjgNFL Paper: https://t.co/Qs2mEdgCTP GitHub: https://t.co/bvtqM9GMa9
6
60
278
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 https://t.co/VmfjfuxDgR
1
6
39
Excited! VideoPrism-Base/Large are publicly available now: https://t.co/g5BNiA5O05 Check it out if you need a versatile video encoder for video-language or video-native tasks. Feedback appreciated!
github.com
Official repository for "VideoPrism: A Foundational Visual Encoder for Video Understanding" (ICML 2024) - google-deepmind/videoprism
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 https://t.co/vAVqXo8g4j
0
6
21
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 https://t.co/imLrPYAnEk. We are also working towards adding more support into the repository, please stay tuned.
github.com
Official repository for "VideoPrism: A Foundational Visual Encoder for Video Understanding" (ICML 2024) - google-deepmind/videoprism
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 https://t.co/vAVqXo8g4j
0
1
8
Introducing our latest work Video Creation by Demonstration, a novel video creation experience. Paper: https://t.co/YZFCLKj5aM Project: https://t.co/o9inp7qScE Huggingface:
huggingface.co
0
8
34
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
In generative modeling, tokenization simplifies complex data into compact, structured representations, creating a more efficient, learnable space. For high-dimensional visual data, it reduces...
0
4
11
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 → https://t.co/MnfeIMAohS
22
66
396
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 https://t.co/vAVqXo8g4j
35
214
834
Google presents Video Instruction Tuning Distilling Vision-Language Models on Millions of Videos paper page: https://t.co/MsoJ6GMhGq Experiments show that a video-language dual-encoder model contrastively trained on these auto-generated captions is 3.8% better than the
0
38
181
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!
3
61
397
📢 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! 🏆 https://t.co/asTC5oscJh
#VLAR #ICCV2023 #AI
1
20
17
VideoGLUE: Video General Understanding Evaluation of Foundation Models paper page: https://t.co/Y97nZAXGm9 We evaluate existing foundation models video understanding capabilities using a carefully designed experiment protocol consisting of three hallmark tasks (action
0
23
108
~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!
The Multi-Agent Behavior Workshop (MABe) will take place @CVPR 2023 in Vancouver! We are featuring fantastic speakers @SiyuTang3, @georgiagkioxari, @tlandgraf, @Wei_ZHAN_, @sabinehauert, & Ben Sapp; with panel & poster discussions 👇 Call for papers!📢 https://t.co/pWa7OoIPHF
0
6
19
Bard is now available in the US and UK, w/more countries to come. It’s great to see early @GoogleAI work reflected in it—advances in sequence learning, large neural nets, Transformers, responsible AI techniques, dialog systems & more. You can try it at
gemini.google.com
Meet Gemini, Google’s AI assistant. Get help with writing, planning, brainstorming, and more. Experience the power of generative AI.
We're expanding access to Bard in US + UK with more countries ahead, it's an early experiment that lets you collaborate with generative AI. Hope Bard sparks more creativity and curiosity, and will get better with feedback. Sign up: https://t.co/C1ibWrqTDr
https://t.co/N8Dzx1m0fc
27
118
711
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
57
169
2K
Attending #ECCV2022? Drop by the poster session 1.A (poster 084) about VPT! 🔥❄️ w/ @CornellCIS, @BelongieLab and @MetaAI !
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) 👇
1
5
23
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.
4
32
182
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) 👇
2
6
22
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: https://t.co/QV1KynfVkR Challenge: https://t.co/deeqxcf61L
3
20
87