True Positive Weekly #5

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

Hey! In this issue: machine learning to play Super Mario, a collection of GPT-3 demos and articles, MIT released a deep fake video about failed Apollo 11 mission, a fast vector similarity search library from Google, a talk on machine learning and causal difference and more.

• Machine learning to play Super Mario Bros (video, paper, code)

• Awesome GPT-3 is a collection of demos and articles about the OpenAI GPT-3API

• In this deepfake video from MIT, President Nixon is announcing NASA Apollo 11 disaster (the speech that was written but never read by him)

• The dangers of trusting black-box machine learning

• Facebook released DE⫶TR: end-to-end object detection with Transformers (blog, code, paper)

• Symbolic mathematics finally yields to neural networks

• A talk on machine learning and causal inference by Susan Athey, Stanford

• Self-supervised representation learning in natural language processing

• ScaNN (Scalable Nearest Neighbors) from Google: a method for efficient vector similarity search at scale (blog, code, paper)

---

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