New Delhi: A Google-funded open source voice recognition program- Sirius- promises the possibility to take over the existing virtual assistants with better capabilities.
Sirius is an open source program developed by researchers at the University of Michigan's Clarity Lab and is similar to Apple's Siri or Google's Google Now voice recognition application.
Apart from Google, the Defense Advanced Research Projects Agency, the US military's research wing, and the National Science Foundation are backing the open-source program, a report on Motherboard stated.
The idea of having an open-source program for virtual assistant app is to let anyone contribute to the program on GitHub. As it has been released under a BSD license, it will be completely free for anyone to use or distribute.
One of the researchers who leads the project, Jason Mars, describes Sirius as a Linux-like version of Siri. Sirius is currently only being tested on Ubuntu desktops, but it could soon be brought on other OS on phones and devices.
What makes Sirius different from its counterparts is that it has the ability to answer a question about any picture that you capture and feed to the assistant. Siri can't do that, however, Sirius isn't as elegant as Siri because it is a patchwork of other open source projects which are stitched together to boost Sirius' capabilities.
Sirius uses Carnegie Mellon University's Sphinx program in combination with Caffe, an open source deep neural network platform to recognise speech. To understand images, it uses OpenCV's SURF program and image database, which IBM's Watson also uses. To actually answer questions, Sirius uses a system called OpenEphyra, also developed by researchers at Carnegie Mellon.
One needs to download all of these supporting programs to make Sirius work. The team behind Sirius is running a tutorial on how to make Sirius work, as well as presenting a paper, next week at the 2015 ASPLOS conference for programming languages.
Comments
0 comment