Chao Xing 2016-04-25

From cslt Wiki
Revision as of 00:57, 25 April 2016 by Xingchao (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Last Week :

1. Work done for DNN-DSSM model.

  Get 2 point of results :
  • DNN-DSSM can not work well for word-level grams.
  • DNN-DSSM dostn't work well for similar question task.

2. Work done for single layer CNN-DSSM model.

  Got reasonable results  in DSSM-Report

3. Test for mixture data CNN-DSSM max-margin model.

  Can not achieve good results.

This Week:

1. Test for multi-layer CNN-DSSM model, both max-margin or promote loss.

2. Need some talk.