Unsupervised MT, pre-trained language models, common sense inference datasets, meta-learning, robust unsupervised learning, understanding representations, clever auxiliary tasks, inductive bias and more
"We’ve discovered that the gradient noise scale, a simple statistical metric, predicts the parallelizability of neural network training on a wide range of tasks. Since complex tasks tend to have noisier gradients, increasingly large batch sizes are likely to become useful in the future, removing one potential limit to further growth of AI systems. More broadly, these results show that neural network training need not be considered a mysterious art, but can be rigorized and systematized."
Current state of collaboration around building and deploying models, tension points that potentially arise, as well as practical advice on how to address these tension points.
For situations where scenes don't contain one main object or a simple scene, object detection can be used to improve the performance of computer vision algorithms. Comes with examples from the retail industry.