I just read a great post here. I am curious about content of "An example with images" in that post. If the hidden states mean a lot of features of the original picture and getting closer to final result, using dimension reduction on hidden states should provide better result than the original raw pixels, I think.

Hence, I tried it on mnist digits with 2 hidden layers of 256 unit NN, using T-SNE for dimension reduction; the result is far from ideal. From left to right, top to bot, they are raw pixels, second hidden layer and final prediction. Can anyone explain that?

enter image description here

  • $\begingroup$ How good is your network at performing the task? $\endgroup$ Jul 21, 2016 at 14:30
  • $\begingroup$ @FranckDernoncourt, Thanks for your replay. The accuracy is around 94.x%, which is able to distinguish most of digits. Would that be a reason to cause bad performance on dimension reduction? $\endgroup$
    – Hanyu Guo
    Jul 21, 2016 at 19:14


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.