Project Page: https://
cs.stanford.edu/~xtiange/proje
cts/gaussianatlas/
…
Paper: https://
arxiv.org/abs/2503.15877
MULTIMODAL AI
-
Stanford’s Gaussian Atlas: 3D Scene Representation Breakthrough
By
–
-
2D Diffusion Models Repurposed for 3D Object Generation
By
–
A team led by Dr. Li Fei-Fei @drfeifei found that pre-trained 2D diffusion models can be repurposed for 3D object generation. Their proposed GaussianAtlas enables fine-tuning of 2D diffusion models to generate 3D Gaussians. pic.twitter.com/D2mvz52C08
— 机器之心 JIQIZHIXIN (@jiqizhixin) 21 mars 2025A team led by Dr. Li Fei-Fei @drfeifei found that pre-trained 2D diffusion models can be repurposed for 3D object generation. Their proposed GaussianAtlas enables fine-tuning of 2D diffusion models to generate 3D Gaussians.
-
xAI prepares Grok iOS app with new image editing feature
By
–
xAI is preparing grok iOS app for image editing with a new option being added to the attachment menu. Earlier, image editing option was only available for images generated by Grok. h/t @whylifeis4
-
InfiniteYou: Infinite Image Generation Model by ByteDance
By
–
Paper: https://
arxiv.org/pdf/2503.16418
Project Page: https://
bytedance.github.io/InfiniteYou/
Code: https://
github.com/bytedance/Infi
niteYou
…
Model: https://
huggingface.co/ByteDance/Infi
niteYou
…
Demo: https://
huggingface.co/spaces/ByteDan
ce/InfiniteYou-FLUX
… -
ByteDance InfiniteYou: Identity-Preserved Image Generation Tool
By
–
ByteDance Intelligent Creation released InfiniteYou—upload one photo, and you can appear in infinite universes. Its high-fidelity, identity-preserved image generation is truly impressive! 🚀 pic.twitter.com/dQyyDDgJF7
— 机器之心 JIQIZHIXIN (@jiqizhixin) 21 mars 2025ByteDance Intelligent Creation released InfiniteYou—upload one photo, and you can appear in infinite universes. Its high-fidelity, identity-preserved image generation is truly impressive!
-
Perplexity AI Enhances Deep Research with Chart Rendering and Code Execution
By
–
BREAKING 🚨: Perplexity is working on an improved Deep Research which can render charts and use code execution.
— 🚨 AI News | TestingCatalog (@testingcatalog) 21 mars 2025
This feature may already arrive next week 👀
h/t @AravSrinivas pic.twitter.com/OU6g9iY7amBREAKING : Perplexity is working on an improved Deep Research which can render charts and use code execution. This feature may already arrive next week h/t @AravSrinivas
-
Microsoft Copilot Vision Expanded to Free and Pro Users
By
–
ICYMI: Microsoft Copilot changelog was updated on March 19: Copilot Vision is expanded to free users on Edge and to Pro users on Android in US. Now it can see what's happening on the opened page and has 2 action buttons to summarise it or expand on the topic.
-
A Multi-Tool Generative AI Workflow for Video Production
By
–
> ChatGPT comes up with ideas and makes the script
> ChatGPT generates the prompts for image generation
> Flux AI generates the images based on the prompts
> Kling AI generates the videos based on images
> Elevenlabs API makes the music + voice over
> Creatomate merges into one -
Semantic VAD improves speech-to-speech by detecting natural pauses
By
–
Building smooth speech-to-speech experiences can be tricky, especially handling natural pauses. Our new Semantic VAD in the Realtime API knows if someone’s still thinking or actually done talking, helping you avoid those awkward interruptions!