Stanford researchers recently published results on Neural Network algorithms trained to detect sexual orientation, reporting 91% accuracy when given 5 photos for men and 83% for women.

This article has been highly controversial. Critiques have ranged from concern over the potential for misuse of such technology, to suggestions that the physical signals detected by the Neural Net, and sexual orientation, may have a common cause reaching back to gestation.

What may not be getting enough attention is the source of the training data: The sample data was taken from dating apps of the respective populations. This may introduce an unintended sampling bias in the user's decisions to select certain photos over others in order to appear attractive in the respective dating apps. Further research is needed to rule out such confounding factors.

Apple unveils the new iPhone X with a facial recognition unlock, animated emojis, photo enhancements and other “neural engine” driven features. These are all enabled by the A11 chip designed to support Neural Network based processing.

Physicist Tegmark, who has published a streak of papers applying advanced techniques from theoretical physics to understanding deep learning, discusses the rise in AI. A counter-weight to figures like Elon Musk, advocating for increased regulation.

Uber has a firehose of data growth to manage. Leveraging it effectively is no trivial task. Uber has just released a rundown of Michelangelo, the ML platform that Uber uses for everything from architecting complex ML pipelines to scalable training to deployment.

Uber's Michelangelo

How to use generative adversarial deep learning networks to create new video content from text descriptions.

Latest release by Socher and company continuing, advances on RNN based Q&A and “abstractive” text summarization.

An oldie but goodie. Pandas is still a go to tool for quickly viewing data on tables, joining, grouping by an index and much more.

Syntax for manipulating Pandas DataFrames is not so suspiciously similar to the syntax used when working with the new fundamental data structure in Spark: Spark DataFrames. This is a great place for beginners to get started.

Review of Deeplearning.ai. Andrew Ng’s new deep learning course bringing the teaching clarity in Ng’s lower level courses to more advanced ML material.

Free data science fundamental book including using Notebooks, NumPy (an essential), Pandas, Visualization with Matplotlib, and Machine Learning.