Selected:
Combining next-token prediction and video diffusion in computer vision and robotics
October 18, 2024
A new method can train a neural network to sort corrupted data while anticipating next steps. It can make flexible plans for robots, generate high-quality video, and help AI agents navigate digital environments.
Machine-learning system based on light could yield more powerful, efficient large language models
August 23, 2023
MIT system demonstrates greater than 100-fold improvement in energy efficiency and a 25-fold improvement in compute density compared with current systems.
Learning to grow machine-learning models
April 4, 2023
New LiGO technique accelerates training of large machine-learning models, reducing the monetary and environmental cost of developing AI applications.