“Dongxu Zhang 2015-09-14”版本间的差异

来自cslt Wiki
跳转至: 导航搜索
 
第1行: 第1行:
 
Last Week:
 
Last Week:
  
1. The accuracy of tfidf with 2 layer NN and softmax for classification is higher than 4 layer NN and softmax(79%, 75%).  
+
1. The accuracy of tfidf with 2 layer NN and softmax for classification is higher than 4 layer NN and softmax(79%, 75%). The performance on test set gets best result when training error is around 0.98.  
  The performance on test set gets best result when training error is around 0.98.  
+
  
 
2. Set small lambda of joint training only made the result not that bad. (tried lambda = 0.1)
 
2. Set small lambda of joint training only made the result not that bad. (tried lambda = 0.1)

2015年9月15日 (二) 07:39的最后版本

Last Week:

1. The accuracy of tfidf with 2 layer NN and softmax for classification is higher than 4 layer NN and softmax(79%, 75%). The performance on test set gets best result when training error is around 0.98.

2. Set small lambda of joint training only made the result not that bad. (tried lambda = 0.1)

3. Tried a simple idea on Graph prediction.

This Week:

1. Tried features for softmax. And tried smaller lambda.

2. tune the performance of transE. and check out if the idea works.

3. Try sequence supervised learning. Compare with Viterbi.