http://www.digitaljournal.com/tech-and-science/technology/making-artificial-intelligence-more-private-and-portable/article/507615

Making artificial intelligence more private and portable

Posted Nov 14, 2017 by Tim Sandle
New technology invented at the University of Waterloo is paving the way for artificial intelligence to break free of the Internet and cloud computing, offering a new means of portability.
Engineer Ben Wu manipulating his smartphone as a mouse
Engineer Ben Wu manipulating his smartphone as a mouse
YouTube screenshot
The technology is a form of deep-learning artificial intelligence software developed to fit onto mobile computer chips. This allows artificial intelligence to be used in a range of devices, from smartphones to industrial robots. This portability would enable devices to operate independent of the Internet while using artificial intelligence that performs equivalent to tethered neural networks.
With this, a hosting chip embedded in a smartphone could run a speech-activated virtual assistant and undertake other intelligent features, such as controlling data usage. Other applications include operating drones and surveillance cameras in remote areas.
Commenting on the development, lead researcher Professor Alexander Wong said: "We feel this has enormous potential. This could be an enabler in many fields where people are struggling to get deep-learning artificial intelligence in an operational form.”
The application of stand-alone deep-learning artificial intelligence should trigger lower data processing and transmission costs. Other potential benefits include greater privacy and the ability to the technology in areas where existing systems are impractical.
What is also special about the new development is that conventional deep-learning artificial intelligence, modeled on the human brain and using sets of artificial neurons, requires considerable computational power, memory and energy to function. The new technology, however, has learned to adapt to a harsher operating environment.
To achieve this, the computer scientists placed the artificial intelligence in a virtual environment, and then they progressively and repeatedly depriving it of resources. It was found that the deep-learning artificial intelligence could respond to the altering environment by adapting and changing itself to keep functioning every time computational power and memory were taken away. Essentially the networks made themselves smaller in order to survive in the new resource-deprived environments.
In technology has been presented during the International Conference on Computer Vision in Venice, Italy. A company called DarwinAI has been set up so that the technology can be commercialized.