Siri, Cortana, and to a lesser extent due to its availability on iOS, Google Now – are all virtual assistants to an extent, but limited by availability. Google is apparently backing an open sourced approach to virtual assistants built by the University of Michigan’s Clarity Lab, called Sirius which is taking an open source approach to building the next virtual assistant.
The project, which the team says will eventually run on Mobiles as well as Desktops, has drawn some attention from other big-name players as well as Google, including the Defense Advanced Research Projects Agency (DARPA), the US military’s research wing, and the National Science Foundation, so there’s a lot of interest here.
The Sirius project is headed up by Jason Mars who describes Sirius as a Linux-like version of Siri, which is more configurable because of its open source nature.
With Windows, we don’t have access to change Windows into whatever we want it to be. With Linux, you have your own open-source operating system that you can customise to be whatever whatever you want. And so what we tried to do with Sirius is to provide an open-source, common infrastructure anyone can customise and deploy and have freedom to do whatever they’d like, but have a starting point that is as intelligent as what you would find in these proprietary environments.
Sirius has a number of components which make it a better fit than some of the existing virtual assistants. Sirius already allows you to feed it a picture of say the Eiffel tower and ask questions about it – How old is this?
Sirius isn’t quite ready for primetime however. At this stage you can actually use Sirius, but it involves installing a number of components from various sources, but there is a download suite you can use which makes finding and installing all these sources a bit easier.
But even still, just having the components won’t have you up and running, with a bit of work still needed to get it running. Luckily, the team will be running a tutorial on how to get these components working together when they present a paper on Sirius at the Arhcitectural Support for Progamming Languages and Operating Systems (ASPLOS) 2015 conference currently running.
Will we see Sirius working with Google Now in the future or at least some component integrated into Android in the future? It’s still early days, but it’s an intriguing project nonetheless.