Η Microsoft φέρεται να μετατρέπει τα εργαλεία που χρησιμοποιούν οι ερευνητές της σε open source για να επιταχύνει την πρόοδο της τεχνητής νοημοσύνης, με μια ευρύτερη team developers.
The company shows that it finally made sense, following a tactic used by Google. Open Sourcing is being called upon to accelerate the development of CNTK, a deep learning toolkit developed by Microsoft researchers.
Xuedong Huang, the company's chief scientist, said he and his team want to achieve faster improvements in how well computers can understand speech, and perceive pictures who "see."
So they assigned a problem group to a group of volunteers to solve it on their own. The effort paid off.
In internal testing, according to Xuedong Huang, CNTK proved to be much more effective than the other four popular computing tools used by the company's developers to create models deep learning that are capable of recognizing speech and image.
"The CNTK toolbox is just outrageously more efficient than anything we've ever seen," Xuedong said.
So they decided to make the tools they use, available through open source licenses to all interested researchers and developers.
https://github.com/Microsοft/CNTK