News for hackers. Stuff that matters.
Re:Getting Deep Speech to Work in Mandarin by amake on Monday February 15, 2016 @11:23PM (#11107869)
> Uh, movie subtitles?

Movie subtitles are very rarely actual transcriptions of what is spoken; they are instead summaries, edited for brevity and quick comprehension.

I don't know much about what kind of corpora is required for training this kind of model, but subtitles don't seem appropriate.

Re:Getting Deep Speech to Work in Mandarin by wibr on Tuesday February 16, 2016 @04:00AM (#11108650)
In Mandarin they are 99% exact transcriptions.
Re:Getting Deep Speech to Work in Mandarin by taneq on Tuesday February 16, 2016 @12:40AM (#11108108)
Is that necessarily a bad thing? Maybe if you're transcribing text for dictation, but for many voice recognition uses, you would be happy with text conveying the semantic content of the spoken text.
Re:Getting Deep Speech to Work in Mandarin by amake on Tuesday February 16, 2016 @12:52AM (#11108138)
This is really not my field, but my suspicion is It Just Doesn't Work Like That.
Re:Getting Deep Speech to Work in Mandarin by taneq on Tuesday February 16, 2016 @01:25AM (#11108248)
Not my field either, but I wouldn't have thought that mere word proximity would lead to something as interesting as word embeddings either, and they seem to be a general purpose "input token -> semantic meaning" mapping, so who knows? (Seriously, if someone does know, tell!)
Re:Getting Deep Speech to Work in Mandarin by oergiR on Tuesday February 16, 2016 @02:43AM (#11108427)
To train a speech recogniser on subtitles, you need to be careful to allow for words missing from the transcriptions. But it has been done: goo.gl/Syx27j