Here is a random walk through some of the 2017 year-end summaries of what’s happening with AI, machine learning and deep learning.
WildML
Denny Britz writes a regular newsletter on AI, Deep Learning, and NLP called WildML and this is how he summed up the year:
Facebook describes progress on data centers and connectivity. And in that series, they add a post about how they are using AI to build more immersive user experiences including better translations, style transfer, and 360 video:
Google DeepMind
The DeepMind reflects on their next big Go-playing system AlphaGo Zero as well their work on Parallel WaveNet, which generates even more natural sounding human voices, AI ethics, and delivering real-world benefits to organizations like NHS.
Google Brain
On behalf of the Google Brain team, Jeff Dean discusses both infrastructure work as well as practical applications:
- AutoML, a system which solves new machine learning problems automatically without human experts designing a specific neural network architecture
- TacoTron 2, which also generates very realistic human sounding voices
- The mind-bending work on learning index structures that can generate more compact data representations and support faster queries compared to traditional data structures such as B-trees, hash tables, and Bloom filters.
- Their ongoing work on the infrastructure to support machine learning: TensorFlow, TPUs, and TPO pods
Udacity
Udacity had quite a year, and machine learning, self-driving cars, robotics played a big role. Did you know they have a Flying Car Nanodegree program? I am not making that up.
Databricks
More and more machine learning and AI is happening in Databricks clusters with Apache Spark. And starting in 2018, the Spark Summit is now the Spark + AI Summit.
https://databricks.com/blog/2018/01/03/databricks-and-apache-spark-2017-year-in-review.html
NIPS 2017
This is one of the biggest AI research conferences, and Brown University student David Abel compiled an epic 43-page field guide in case you missed the conference.
AI Hubs—Or, In Which Cities is the Action Most Concentrated?
You know about the Bay Area and Toronto. Here are eight other cities that are turning out to be the leading clusters of AI research and development.
Artificial Lawyer
A site dedicated to covering “tech that performs work in a legal sense” interviewed a set of subject matter experts to get a roundup of 2017’s biggest hits and predictions for 2018.
