What’s even more fascinating is the fact that today’s voice recognition is not just limited to one single language – ” English” – but can recognize up to 119 languages from across the globe, with various accents, tone and can achieve ~95% accuracy rate. Thanks to artificial intelligence and neural networks, today’s computers can’t only understand English and its 1 million+ word and phrases but can even adapt to your tone, speech pattern and accent. The only thing that kept voice recognition from becoming the next dominant computing feature was unreliability.īut today things are changing in the landscape of voice recognition’s multi-verse. So, basically, they couldn’t differentiate between “broken legs” and “beacon eggs.” The Modern Speech Recognition Even nearly 30 years later that number had grown to only around twenty thousand, which may seem like a lot but remember that the English language has over one million words on top of that early software couldn’t predict what words you were trying to say by using context. Some of the first systems, like Audrey from the 1950s, could only recognize about 10 words. 1952 when Bell Laboratories first designed the “Audrey” system to Fast forward twenty-five years and voice recognition technology is kind of VR (virtual reality) where someone is always claiming that they’ve made a significant breakthrough and it turns out to still be kind of crap.Įarly forms of voice recognition had very limited vocabularies.
0 kommentar(er)
