After releasing a statement on its artificial intelligence development efforts, Facebook’s AI looks like it’ll kick Google’s Smart Reply right to the curb. Chief Technological Officer Mike Schroepfer took to to the Facebook newsroom to announce the latest in Facebook’s AI, including not only a function that allows Facebook to become accessible to the blind, but the ability for the AI to understand language, but also recognize objects, make predictions and make planning.
Among the most surprising revelations were just how accurately Facebook’s AI an not only understand and recognize language, but it can also recognize images. The system, called Memory Networds or MemNets, can read and answer questions about short texts. Similarly to Google’s new Smart Reply Tool, the system can read text, understand, and devise a way to respond. But unlike Google’s Smart Reply Tool, it cannot recognize images. Facebook’s AI allows users to even ask it what’s in a photo, and the artificial intelligence answers simply and accurately.
Additionally, Facebook’s AI will have object detection integrated into its systems. The new upcoming system would segment images 30% faster than any other artificial intelligence using 10 times less data.
All things point to the fact that Facebook is measuring up to both Google and Apple in terms of artificial intelligence, a tool that has become increasingly smart and useful for all consumers. It’s yet to be seen what the competing companies will bring to the table next.
