Baidu has demonstrated that a single deep learning voice system could learn to reproduce thousands of speaker identities, with less than half an hour of training data for each speaker. This capability was enabled by learning shared and discriminative
You are free to comment on this discussion in any way you feel is appropriate. If you choose to use to use any language which our editors feel is vulgar -- by their standards -- your comment may be tagged "Crude or Lewd" and may be filtered out of the discussion by those who prefer not to read that sort of thing. If you know you have entered something which will cause your comment to be tagged, we ask that you tag it yourself to save us the time. We do encourage everyone to be civil and not make rude attacks on other people in the Forum. We don't censor out those remarks, but few people enjoy reading them and we would like participation in our Forums to be a pleasant experience for everyone. And, by concentrating on what is said instead of who is saying it, even those who may disagree with you will be more likely to consider your opinions valid.