Baidu proves once again that it can play cognitive open source at the world class level of a Google or a Facebook or a Microsoft as it announced the availability of four core speech technologies. These are Long Utterance Speech Recognition, Far-Field Speech Recognition, Expressive Speech Synthesis and Wake Word.
The speech APIs join a growing list of public Baidu AI resources, including PaddlePaddle, its deep learning platform, and technologies for cognitive applications including: facial recognition, optical character recognition, and natural language processing. Baidu had begun releasing speech APIs in 2013, and its previous speech-oriented APIs included speech recognition, speech synthesis, and user-defined semantics.
Baidu claims that the number of developers using their speech systems has grown from 10,000 in 2014 to a projected 140,000 in 2016.
Read more at Yahoo Finance:Share