True Positive Weekly #12

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

Hey, in this issue: a curated list to track current scary usages of AI, who is responsible when autonomous systems fail, what businesses need to know about model explainability, turning any CNN image classifier into an object detector, how machine learning finally reached the enterprise, and more.

• Awful AI: a curated list to track current scary usages of AI hoping to raise awareness to its misuses in society

• A new training method exceeds GPT-3 performance with 99.9% fewer parameters (code, data)

• AI is an ideology, not a technology

• Who is responsible when autonomous systems fail?

• Generating cooking recipes using TensorFlow and LSTM Recurrent Neural Network: a step-by-step guide

• [Podcast] What businesses need to know about model explainability

• Google Brain’s SimCLRv2 achieves new SOTA in semi-supervised learning (code)

• Hummingbird compiles trained ML models into tensor computation for faster inference

• Turning any CNN image classifier into an object detector with Keras, TensorFlow, and OpenCV

• The secrets of small data: how machine learning finally reached the enterprise

• [Announcement] Machine Learning Engineering, the book by Andriy Burkov is now released on Leanpub and Amazon (paperback, hardcover)

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

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