*Now that there's a ton of voice traffic on the Internet, major-league search-engine techniques can be brought to bear on it.
*That's a little creepy.
http://radar.oreilly.com/2010/04/big-data-shakes-up-the-speech-industry.html
(...)
"Having speech technologies in the cloud lets Google quickly iterate and push enhanced speech engines on a regular basis. More importantly, their speech engines learn and get trained using real data from their many interconnected services. Speech engines typically rely on both language and acoustic models. Language models are statistical models of word sequences and patterns. Cohen pointed out that their language models use data collected from web searches, giving them access to an ever growing corpus that few can match (230 billion words collected, refined to a vocabulary of the million most common words).
(((So, what kind of mixmaster, guitar-pedal effects can one get out of a "spoken language model," one wonders. Like, if you had the comprehensive archive of all phone calls made to French cops, could you use one to call a French cop, even if you didn't speak French? Maybe you could just input a few factual parameters about the incident and have it generate French for you.)))
(((Then there are the obvious machine-generated phone-sex applications. I was writing sci-fi about those twenty years ago.)))
"Cohen disclosed that some of the more recent acoustic models they're evaluating are built using unsupervised machine-learning algorithms. (These are speech algorithms trained on recorded speech that haven't been transcribed by hand.) (((Oh dear.))) While he coyly avoided explaining how an accurate system can be built from unsupervised techniques, it's likely they use data from their 411 service....
