Google gives just 4.9% word error rate in Speech Recognition
At Google I/O 2017, Google’s CEO Sunder Pichai proudly announced that the error rate in speech recognition of the company has dropped to an all-time low of 4.9%, which stands as a much better figure than 23% in 2013.
Pichai also shared that the main reason for such a slash in word error is due to the usage of ‘Neural Network’ technology, which amplifies the performance of speech recognition considerably. Google’s products using speech recognition have become considerably preferable even in noisy areas, where they can easily pick up speech and words.
These techniques of amplified speech recognition have certainly made room for an error-free system that can rudimentarily comprehend every word of speech. With Google reaching an all-time word error rate, other companies have also fastened their seatbelts to match the consumer base of this tech giant and maybe in near future, we may witness some phenomenal in this field of development.
Image Source: zdnet3.cbsistatic