Survey data gathered voluntarily from homeless youth, as well as behavioural theories and observations of previous interventions, were used to build a computational model of the interventions.
These pathology images were subdivided into 40,000 smaller samples; 30,000 of these samples were used to 'train' the software, the remaining 10,000 were used to test accuracy.
Researchers from the company's Asia and US labs said their system achieved human parity on a commonly-used test set of news stories -- called "newstest2017".