Slightly OT, but in school I remember a teach talking about how good voice to text has gotten but was stuck, stating it was roughly about a 70% accuracy, but that we were pretty much at a wall there and haven't moved much in the last decade.
Has there been any major advances in voice recognition other than just growing your speech corpus?
I've actually taken to dictating most of my texts on Android. It's faster for me than typing or swype-style keyboards and it's really very accurate until I need to use a strange proper name it doesn't understand. I'd say it easily gets about 90% of what I say, and it figures out the correct context for words like "there" "their" "they're" and "to" and "too" so far completely correct.
edit thought I'd add this. I just had a conversation with a united agent and they probably understood less than 70% of what I was saying. it was beyond frustrating I wish that I was actually talking to my phone instead.
How do you write text messages in private and not have your messages overheard? I think I would be quite self concious that I'm talking into my phone and not directing it at anyone (I rarely use Siri for this reason)
Yea I'm with you. I do voice searches/typing CONSTANTLY when I'm at home, particularly on weekend mornings when I'm checking my schedule/weather/texting friends to organize my activities for the day while getting ready. Barring the occasional query when I'm on the sidewalk and not too near anyone, I don't really use them outside much.
well, I'll type it then, of course. But I'm usually locked away working in an office or working at home most of the time and not around too many people.
I think it's been mostly throwing hardware at the problem. Server hardware, backed by terabytes of context data.
I remember back in the day, most of the errors were the exact same errors that a human would make. Even you and I only really hear 95% or so of the words someone says. But we can fill in the rest from the context. It always amazes me when I dictate some sentence to my phone, and at the end I see one of the words change to another that sounds almost identical but makes much more sense in the entire context of the sentence.
Has there been any major advances in voice recognition other than just growing your speech corpus?