A month ago, Microsoft's dialogue speech recognition technology achieved a breakthrough in the word error rate (WER) as low as 6.3% in the industry standard Switchboard speech recognition benchmark test, creating the lowest error rate in the field at that time.

Recently, Microsoft has further reduced the word error rate to 5.9%, achieving for the first time the performance equal to that of professional stenographers and better than most people.

Microsoft's speech recognition technology can set records again and again, thanks in large part to CNTK (ComputaTIonal Network Toolkit), an open source tool.

Microsoft today released a "cognitive toolkit" for speech recognition technology

The system was originally established to study speech applications, and later developed into a localized deep learning system for Microsoft. The CNTK toolkit was open sourced on GitHub a year ago, and currently speech recognition including Microsoft AI personal assistant Xiaona and HoloLens is based on CNTK. The biggest difference between CNTK and other open source software is that it can do large-scale, distributed machine learning, while ensuring powerful performance.

Today, the toolkit has been updated to add a beta version known as the "cognitive toolkit."

The research on speech recognition can be traced back to a related project funded by the DARPA (Defense Advanced Research Projects Agency, the US Department of Defense Advanced Research Projects Agency, mainly dedicated to the research, development, and application of US defense high-tech) in the 1970s. In the decades since, more and more research institutions and large companies have joined it.

"This breakthrough is the result of the continuous accumulation of speech recognition technology over the past two decades," said Geoffrey Zweig, a researcher at Microsoft's Language and Dialogue Research Group.

Microsoft today released a "cognitive toolkit" for speech recognition technology What is the concept of Microsoft's 5.9% word error rate?

In the industry standard Switchboard speech recognition task test, the human control group (composed of professional stenographers) transcribes the dialogue speech into text. The minimum word error rate currently recorded is 5.9%, which means the speech of Microsoft's speech recognition system Recognition ability is already higher than that of most people in the world and it is equal to that of professional human beings, creating a new world record.

This milestone breakthrough in speech recognition will have a profound impact on consumers and commercial products, because speech recognition technology can significantly enhance people's daily computing experience. These products include entertainment devices like XBOX, productivity tools like Microsoft Cortana, and personal artificial intelligence assistants that enable real-time voice-to-text transcription.

Although the word error rate of the Microsoft team has reached the same level as 5.9% of human professional stenographers, this does not mean that the computer can perfectly recognize every word. If one day the computer can perfectly recognize "blue skinny" and "shiitake", then voice recognition technology will reach a new height.

Voice Keychain

voice keychain:

Ningbo AST Industry Co.,Ltd has 17 Years Experiences to produce the Voice Keychain,musical keychains,keychains,led keychains and so on.we can assure you of competive price,high quality,prompt delivery and technology supporting.



1. Voice Keychain Application:

The Voice Keychains are usually used for Celebration,Promotion,Birthday,Holiday,Award Ceremony and so on.

2. Classification:

There are Recordable keychain and Musical keychain
Recordable keychain--we can record our own message and playback it
Music Keychain--The audios will be customized.The clients provide the audios,the we programme the audio into the voice keychain.then it will play the customized audio


3.Sound Quality:Clearly and Loudly


4.Play activated:Press the play button to play the message


5.Can be printed your logo on it


6.Certification:CE ,Rohs


7.Export to:

USA, UK,Canada,Germany, Turkey,Russia,Poland,Switzerland,Netherland ,Frence Hungary ,Australia,New Zealand, Brazil, Columbia,Argentina,Thainland,Singapore , Malaysia and so on.



Voice Keychain, Musical Key Chain, Voice Recorder Keychain

AST Industry Co.,LTD , https://www.astsoundchip.com