Private, Portable Deep Learning

More mobile deep machine learning:

New Technology Makes Artificial Intelligence More Private and Portable 
University of Waterloo News  by Matthew Grant

Researchers at the University of Waterloo in Canada say they have developed new deep-learning artificial intelligence (AI) software that could enable other AI technologies to break free of the Internet and cloud computing. The technology is compact enough to fit on computer chips that can be used in mobile devices and industrial robots, which would enable devices to operate independently of the Internet while using AI that performs almost as well as tethered neural networks. The researchers say the technology could lead to much lower data processing and transmission costs, greater privacy, and use in areas where existing technology is impractical due to cost or other factors. The researchers based the new device on evolutionary forces in nature to make the AI more efficient by placing it in a virtual environment, and then progressively and repeatedly depriving it of resources. The deep-learning AI adapts and changes itself to keep functioning each time computational power and memory are removed.  ... " 

European Commision on Digital Platforms

In the Economist (registration required) on pricing and Google in the online world.   Are Google (and other large digital platform holders) preventing competition by buying up potential competitors early with their cash?  Also the data resources needed?   In particular with regard to online platforms. Very difficult to forecast the future of small startups.  (Been asked to look at the forecasting aspect of this, so some of the tabs will be filled out later)