The
Clarity Lab group, based at the University of Michigan, offers an open source solution for a personal computer assistant. Their program recognizes voice and looks for answers to requests, just like Apple’s Siri, Google Now Now from Google, and similar solutions from Microsoft and Amazon. The project called Sirius can even more than its counterparts - for example, you can provide it with a picture and ask a question about it.
The project is sponsored by Google, DARPA, the US military and the US National Science Foundation. It is distributed under
the BSD license , which means complete openness and gratuity for everyone. It means that any start-ups will be able to implement such functionality in their projects and devices.
The principle of the system
')
But for this, they will have to figure out the rather complex system that makes up the project
laid out on GitHub . The system requires additional separate programs -
Sphinx , Kaldi, Protobuf, Speeded Up Robust Features (
SURF , works on the basis of OpenCV). At the ASPLOS 2015 conference, developers will present both their research work on the project and the
tutorial on installing and working with it.
Among enthusiasts have already been attempts to create an analogue of Siri - this, for example,
JuliusJS , a library for speech recognition. But so far no noticeable results from these projects are visible.