Search the web
Sign In
New User? Sign Up
icsi-speech-tools
  • Members Only
  • Post
  • Files
  • Photos
  • Links
  • Database
  • Polls
  • Members
  • Calendar
  • Promote
  • Groups Labs (Beta)
? Already a member? Sign in to Yahoo!

Yahoo! Groups Tips

Did you know...
Message search is now enhanced, find messages faster. Take it for a spin.

Best of Y! Groups

   Check them out and nominate your group.
Having problems with message search? Fill out this form to ensure your group is one of the first to be migrated to the new message search system.

Messages

   Messages Help
Message #
Search:
Advanced
Why are there positive posteriori probabilities when I use MLP featu   Message List  
Reply | Forward Message #154 of 162 |
When I use MLP as my HMM features, the POSITIVE posteriori
probabilities always occurred. The POSITIVE posteriori probabilities
will break the decoding rule and result in bad performances. I don't
know whether you found the same problem, and what can I do to avoid
this? Thank you very much!
Thu Oct 30, 2008 7:26 am

Show Message Option
qingqing.zhang
Offline Offline
Send Email Send Email

Forward 		Message #154 of 162 |
Expand Messages Author Sort by Date

When I use MLP as my HMM features, the POSITIVE posteriori probabilities always occurred. The POSITIVE posteriori probabilities will break the decoding rule... 		qingqing.zhang
Offline Send Email 		Oct 30, 2008
7:26 am

... I'm not sure I understand. If you're using softmax as the output layer of the MLP, then all the values should be between 0.0 and 1.0. If you're not using...
Adam Janin
ajanin
Offline Send Email 		Oct 30, 2008
4:58 pm

Dear Adam, I am so sorry that I did not explain the problem clearly. Actually, we found that MLP has the much larger dynamic range than PLP has (In the log...
qingqing.zhang
Offline Send Email 		Nov 3, 2008
3:26 am

There is another problem: When we use MLP feature on Mandarin CTS, the Frame Error Rate of training data was up to 60% after NN training. Does this a little ...
qingqing.zhang
Offline Send Email 		Nov 3, 2008
7:37 am

... I haven't run Mandarin CTS myself, so maybe somebody else on the list can answer in more detail, but for English, high 60s for frame ACCURACY (not error)...
Adam Janin
ajanin
Offline Send Email 		Nov 3, 2008
7:23 pm

From your message, I assume you're using the MLP outputs as features to a GMM (so-called "tandem") rather than using the posteriors directly (so-called...
Adam Janin
ajanin
Offline Send Email 		Nov 3, 2008
7:21 pm
< Prev Topic  |  Next Topic >
Message #
Search:
Advanced
Copyright © 2009 Yahoo! Inc. All rights reserved.
Privacy Policy - Terms of Service - Guidelines - Help