Read the full post here.
Ryan Prenger and Tony Han, research scientists at Baidu’s Silicon Valley AI Lab, have posted a Tech Note about how SVAIL enabled the Baidu Deep Speech system to recognize Mandarin in about sixty days, an ambitious project. They write:
…In just a few months, we produced a Mandarin speech recognition system with a recognition rate better than native Mandarin speakers, in some cases. The biggest change we had to make for our Deep Speech system (originally developed in English) to work in Mandarin was increasing the size of our output layer to accommodate the larger number of Chinese characters.
In the Tech Note, they provide details on what the team did to adapt the system to Mandarin and how the Baidu end-to-end learning approach made the project achievable.
See also our SVAIL GitHub Blog.
SVAIL Tech Notes are written by engineers for engineers on topics related to AI technologies, techniques, tips and trends. Previous issues:
Deploying Deep Neural Networks Efficiently, by Chris Fougner
Optimizing RNN Performance, by Erich Elsen