A monthly roundup of news about Artificial Intelligence, Machine Learning and Data Science. This is an eclectic collection of interesting blog posts, software announcements and data applications I've noted over the past month or so.
Open Source AI, ML & Data Science News
Intel open-sources NLP Architect, a Python library for deep learning with natural language.
Gym Retro, an open source platform for reinforcement learning research on video games.
Facebook open-sources DensePose, a toolkit to transform 2D images of people into a 3-D surface map of the human body.
MLflow, an open source machine learning platform from Databricks, has been released.
Industry News
In a 12-minute documentary video and accompanying Wired article, Facebook describes how it uses Machine Learning to improve quality of the News Feed.
In the PYPL language rankings, Python ranks #1 in popularity and R is #7; both are rising.
Google announces its ethical principles for AI applications, and AI applications it will not pursue.
Wolfram Research launches the Wolfram Neural Network Repository, with implementations of around 70 neural net models.
Google Cloud introduces preemptive pricing for GPUs, with discounts compared to GPUs attached to non-preemptable VMs.
Conversational AIs conducting conversational full-duplex conversations: Microsoft Xiaoce and Google Duplex.
Microsoft News
Microsoft's head of AI research, Harry Shum, on "raising" ethical AI.
Microsoft has acquired Semantic Machines, a startup focused on conversational AI.
Microsoft is developing a bias-detection tool, with the goal of reducing discrimination in applied AI.
Microsoft's Bot Builder SDKv4 further simplifies the process of developing conversational bots.
Cognitive Services Labs, which offers previews of emerging Microsoft Cognitive Services technologies, adds labs for anomaly detection, ink analysis and more.
ML.NET 0.2, a cross-platform open source machine learning framework for .NET developers, has been released.
Microsoft R Open 3.5.0 has been released.
Azure Databricks now provides a machine-learning runtime and GPU support.
Learning resources
A tutorial on visualizing machine learning models with LIME, a package for R and Python.
A visual introduction to Machine Learning, Part II: Model Tuning and the Bias-Variance Tradeoff, with an in-depth and graphically elegant look at decision trees.
Materials for five new AI-oriented courses have been published to the LearnAI Materials site.
A Developer's Guide to Building AI Applications: a free e-book from O'Reilly and Microsoft.
Microsoft Professional Program for Artificial Intelligence, a free, self-paced on-line certification in AI skills.
The Azure AI Lab provides complete worked applications for generative image synthesis, cognitive search, automated drone flight, artistic style transfer and machine reading comprehension.
CVAE-GAN: a new generative algorithm for synthesizing novel realistic images, like faces.
Land cover mapping with aerial images, using deep learning and FPGAs.
Improving medical imaging diagnostics, using the new Azure Machine Learning Python package for Computer Vision.
Find previous editions of the monthly AI roundup here