News for hackers. Stuff that matters.
Re:Andrew Ng on the state of deep learning at Baidu by anko on Monday February 02, 2015 @09:58PM (#8988625)
Andrew is simply saying (if I read him correctly) that if you use generative models you might classify sounds into something other than Phonemes on one layer, and that classification may actually be better than "Phonemes" for recognising sounds. Phonemes are a construct for us to express a concept with language, but might not be the best way to represent a sound in a machine. So why try to write the routines to map sounds to phonemes when you can train it to just learn the categories that help it understand how to recognise the sounds better.
Re:Andrew Ng on the state of deep learning at Baidu by yongjik on Monday February 02, 2015 @10:53PM (#8988768)
Hmm, OK, that sounds much more reasonable. :)