Facebook Diving Deeper into AI With Text & Video Recognition
The second day at F8, Facebook’s annual developer conference, saw quite a few updates from the social networking giant.
Facebook’s Chief Technology Officer, Mike Schroepfer, addressed their “long-term technology investments” in Connectivity Lab, Facebook AI Research and Oculus.
Speaking about Facebook’s AI endeavour, Schroepfer said demonstrated how they have progressed so far. Working with researchers in the field Facebook is tapping a relatively new advancement in technology called Memory Networks “which enables a machine to perform relatively sophisticated question answering.” Not being limited to Q&A, the AI system can mne images and recognize actions like sports in videos, as VentureBeat points out.
“You can really get these systems to understand deep, minute differences,” Schroepfer said.
A paper was published in October last year, through Cornell University, describing a new class of learning models called memory networks. Facebook spoke about it later in November and since then has been making strides in the direction.
It is becoming common practice for tech heavyweights to invest in Artificial Intelligence and Facebook is no exception.
Facebook’s video recognition advancements surface at the wake of revelations made by a startup called Clarifai about technology that surpasses image recognition to video analysis.
Image credit: Facebook