AI & ML
AI Training Efficiency Doubles Every 16 Months
We’re releasing an analysis showing that since 2012 the amount of compute needed to train a neural net to the same performance on ImageNet classification has be
#neural-networks
3 articles
AI & ML
Neural Networks Face New Threat From Adversarial Attacks
Researchers have identified a critical vulnerability affecting neural network policies used in machine learning systems worldwide. Adversarial attacks—carefully
AI & ML
How GANs, Inverse Learning Connect to Energy Models
Researchers have uncovered significant theoretical connections between three foundational concepts in machine learning: generative adversarial networks (GANs),