Researchers tried to build an autonomous scientist with AI. How'd it go?
Researchers decided to see if GPT-4 could generate and test hypotheses without human guidance. What happened?
aimodels-fyi
Meta unveils Emu Video: Text-to-Video Generation through Image Conditioning
The new approach uses "explicit image conditioning" for higher quality videos
Meta unveils Emu Edit: Precise image editing via text instructions
Existing systems struggle to interpret edit instructions correctly. Emu Edit tackles this through multi-task training.
You can predict disease progression by modeling health data in latent space
Forecasting personalized disease progression by modeling clinical data in a latent space
Researchers taught GPT-4V to use an iPhone and buy things on the Amazon app
It's still early, but a GPT-4V agent can navigate smartphone GUIs using a combination of image processing and text-based reasoning.
AI can discover hidden relationships in tabular data
Identifying new classes within unlabeled data sets
LCMs are a new way to generate high-quality images much faster
LCMs achieve similar quality results to LDMs, but in just 1-4 steps instead of hundreds.