Yutian Chen
@yutianc
Followers
1K
Following
861
Media
11
Statuses
216
Senior Staff Research Scientist at DeepMind
Cambridge, England
Joined November 2009
RT @joshwoodward: The wait is over. @GeminiApp is now shipping Veo 3 *globally* for all Pro members!. That means India, Indonesia, all of E….
0
471
0
RT @XingyouSong: Seeing text-to-text regression work for Google’s massive compute cluster (billion $$ problem!) was the final result to con….
0
73
0
RT @HashemGhaili: Prompt Theory (Made with Veo 3). What if AI-generated characters refused to believe they were AI-generated? https://t.co/….
0
4K
0
Also a new generative image model.
Get ready for Imagen 4 🎨 capable of creating richer images, with more nuanced colors, intricate details and superior typography. Tap each photo below to see more. 👀
0
0
4
Excited to share our new generative video model, Veo 3, now with audio!.
Video, meet audio. 🎥🤝🔊. With Veo 3, our new state-of-the-art generative video model, you can add soundtracks to clips you make. Create talking characters, include sound effects, and more while developing videos in a range of cinematic styles. 🧵
1
1
13
RT @sundarpichai: 1/ Today, Veo 2, our state-of-the-art video model, is rolling out to Gemini Advanced + Whisk!. You can create 8s, high-re….
0
341
0
Amazing videos made by creators using our Veo 2 model!.
Creators are bringing their wildest ideas to life with Veo 2. 🎨
2
0
8
RT @labsdotgoogle: This is KITSUNE - a tale of love between two souls. 🦊 💫. Written, created and directed by @henrydaubrez. All shots gener….
0
165
0
Strong debut of image 3 in text-to-image Arena!.
Breaking news from Text-to-Image Arena! 🖼️✨. @GoogleDeepMind’s Imagen 3 debuts at #1, surpassing Recraft-v3 with a remarkable +70-point lead! Congrats to the Google Imagen team for setting a new bar!. Try the best text2image at LMArena and cast your vote! More analysis👇
0
0
3
RT @XingyouSong: Text-to-text regression will be extremely important for science. Why? Skip feature engineering entirely and absorb **all**….
0
120
0
Excited to share our latest and SOTA video generation model, Veo 2! It generates high quality videos up to a 4K resolution. A great collaboration with my amazing colleagues at Google DeepMind.
Today, we’re announcing Veo 2: our state-of-the-art video generation model which produces realistic, high-quality clips from text or image prompts. 🎥. We’re also releasing an improved version of our text-to-image model, Imagen 3 - available to use in ImageFX through
2
5
52
LLMs are good at conversations but not at numerical reasoning without the help of CoT or search. However, we find their embedding is actually pretty useful for numerical regression. We can now extend Bayesian optimization to arbitrary search spaces using pretrained LLMs.
How do we push the boundaries of Bayesian Optimization to handle complex, irregular, or non-numerical search spaces? In Embed-then-Regress, we use LLM embeddings as a universal representation for string inputs, unlocking BO's potential across new frontiers where traditional
0
3
17
RT @NobelPrize: BREAKING NEWS.The Royal Swedish Academy of Sciences has decided to award the 2024 #NobelPrize in Chemistry with one half to….
0
9K
0