True Positive Weekly #46

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

Hey, in this issue: a primer on ML interpretability and explainability; why adversarial image attacks are no joke; a tutorial on composing decision forest and neural network models; how AI is hijacking art history; Reddit’s ML model deployment and serving architecture; and more.

• [The Gradient] Explain yourself - a primer on ML interpretability and explainability

• [Awful] AI is making applying for jobs even more miserable

• Why adversarial image attacks are no joke

• [Tutorial] Composing decision forest and neural network models with TensorFlow

• How AI is hijacking art history

• AI-designed organisms made from frog cells are capable of self-replicating in a way that differs from plants and animals (also check this)

• A whopping 301 newly confirmed exoplanets – discovered with new deep neural network ExoMiner (plus here and here)

• Reddit’s ML model deployment and serving architecture

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