Want to participate in building a world of intelligent personal assistants? The opportunity awaits at SparkingTogether, where researchers, programmers, and companies can contribute features, behavior and knowledge to an online platform, dubbed FIONA, for creating next-gen virtual avatars. FIONA stands for Framework for Interactive Services Over Natural-conversational Agents.
“People sparking together” is how Patricia Lopez, marketing manager at Adele Robots, the robotics startup behind the platform, describes the system. Contributors create code or design that gets wrapped in the FIONA API so that it can be converted into a Spark – which is an application that can become part of the avatar, whether that be its voice, language or a function (NLP, text-to-speech, computer vision, or 3D design, for instance). The company will host a Sparkstore where developers can sell, or freely share, their Sparks with the world, and those interested in using avatars can then combine different Sparks together in the Sparklink environment. Sparkrender is a capability it’s developed for users to post their avatars – which run on Adele Robots’ servers in the cloud – on their websites or mobile apps.
As Lopez sees it, the environment presents the opportunity to get AI moving a lot faster than it traditionally has. It hopes to do so by bringing together the efforts of different parties working in associated fields and letting them contribute their individual developments to improving machine intelligence so that they can be combined with other efforts in a collective fashion. “That could be amazing if companies, geeks, research groups work together — we can get it done faster,” she says.
Adele Robots already is talking with several universities in the U.S. and Europe about the prospect, she says. “The only thing they have to do is wrap their investigations, their developments, with the API we give them so it can communicate with the rest of the platform,” Lopez says. “You can’t just give the market the latest developments being done in research groups [as-is]”.
To what ends will users put these artificial minds? One example, she says, would be an airline, which could combine different sparks to create an online avatar that would deliver information about its flights to users.
One research group the company has talked with is making a Spark that would would give a virtual avatar facial recognition abilities, which could come in handy for online video communications with users. A company could use that for online customer support, tailoring the interaction based on whether a customer is smiling for a woman or man, for instance, or whether or not they are smiling. “Companies are interested to make statistics about what kind of people interact with the avatar,” she says. “It’s related with analytics, too, like what kind of questions the customer asks more, so that you can make some kind of information that is always repeated by the customer automated for the virtual avatar.”
The project is still in beta mode, with some open source Sparks currently available. The first implementation of its technology will be on the website of Spain’s Oviedo University, she says, by year’s end. It will help students with getting the information on their courses that they need.
The service won’t formally launch until middle of next year, by then ideally with plenty of Sparks in its quiver. “The more Sparks that are in FIONA and the more intelligent the Sparks are. the more intelligent the avatar will be,” says Lopez.