True Positive Weekly #27

The most important artificial intelligence and machine learning links of the week

Hey, in this issue: GitHub and OpenAI release an AI pair programmer, random forests and gradient boosted trees are now native in TensorFlow, on assessing the test error of a neural network on unlabeled data, a tutorial on pre-trained transformers with Hugging Face, how to do multi-task learning intelligently, and more.

• GitHub and OpenAI release Copilot: an AI pair programmer

• Random forests and gradient boosted trees are now native in TensorFlow

• [ArXiv] On assessing the test error of a neural network on unlabeled data

• [Tutorial] Pre-trained transformers with Hugging Face

• A new map created by AI reveals hidden links between Milky Way and Andromeda galaxies

• How to do multi-task learning intelligently

• Same or different? The question flummoxes neural networks

• This AI lets you generate new verses from your favorite rappers

Enjoy the newsletter? Help us make it bigger and better by sharing it with your colleagues and friends.

Have a nice weekend! See you next week. — Andriy