Voice assistants have come a long way, and it’s about time they went open source. Researchers in the University of Michigan have done just that, with their free and open source system, known as Sirius. The name, an obvious play on Apple’s Siri, is the first free and open source voice assistant of its kind. It is being demonstrated at the International Conference on Architectural Support for Programming Languages and Operating Systems in Turkey. Companies have become much more serious about voice assistants recently, with Google working on making its Google Now better, while Microsoft is rumoured to be launching Cortana for both Android and iOS.
A paper has been published on Sirius, which states that the voice assistant uses implements core functions, such as image matching, speech recognition, question and answers and natural language processing. The complete paper is linked above.
According to the paper, Sirius is capable of processing voice commands on a mobile device and responding with audible responses for them. Here’s a video demo of Sirius presented by the team from the University of Michigan.
Voice assistants have become an important part of mobile devices, especially with the advent of wearables on the horizon. That said, there is no open source code for these yet. Apple, Google and Microsoft have all designed their versions of digital personal assistants, in the form of Siri, Google Now and Cortana and even though startups have tried to design their own, none have really come close to the big three. In the same way, while Microsoft has shown impressive results with Cortana, neither of the two are anywhere close to Google Now.