Facebook is intensively exploring voice recognition technology, CNBC reports.
The main conduit for Facebook’s work is its Oculus VR headset platform, with users able to activate an AI assistant while wearing the Oculus Rift or Oculus-powered Samsung Gear VR headsets by saying, “Hey Oculus.” They can then do things like search for games in the app store by voice query.
A Facebook R&D team started working on speech recognition technology in earnest in 2014, when the company acquired speech translation app Jibbigo. Last year, Facebook launched a tool that uses speech recognition to automatically generate captions for videos.
Now, Facebook engineers’ aim is to get the speech recognition technology to function offline, with a longer-term goal being to support other languages beyond English. Long, long-term, Facebook product manager Merlyn Deng tells CNBC the aim is “to build a deep semantic understanding of people’s interests, and also of content.”
These sound like ambitious aims, but Facebook has a lot of catching up to do. Google and Apple have ramped up their investments in voice recognition technology in recent months, and Amazon has a trove of data upon which to train its algorithms thanks to the market success of devices using its Alexa AI platform. Of course, there’s greater incentive than ever to push forward in this area, with voice interaction expected to be the central user interface of the Internet of Things.