News

ByteDance Releases USO Model: Breaking the Dichotomy Between Style and Theme in AI Image Generation

ByteDance researchers understand that the growth of AI models relies on data. Therefore, they constructed a large dataset containing approximately 200,000 triplets. Each triplet includes a “style ...
Artnet

Animations, and 3-D Models, and 3,000 Drawings: Inside Google’s Massive Machine-Learning Masterclass on Leonardo da Vinci

What can A.I. teach us about Italian Renaissance polymath Leonardo da Vinci? A lot, as we discovered from a new online retrospective from Google Arts and Culture that’s powered by machine learning.