“140902”版本间的差异

来自cslt Wiki
跳转至: 导航搜索
Lr讨论 | 贡献
 
(相同用户的16个中间修订版本未显示)
第1行: 第1行:
 
== QA system ==
 
== QA system ==
 
* Select the text question (100) and type the class label
 
* Select the text question (100) and type the class label
* read the paper "Building a Question classifier for TREC-Stytle Question Answering System" [http://nlp.stanford.edu/courses/cs224n/2004/may-steinberg-project.pdf]
+
:* i don't know how to label the clsses(hold),maybe the classes is not reasonable.
 +
* read the paper "Building a Question classifier for TREC-Stytle Question Answering System" [http://nlp.stanford.edu/courses/cs224n/2004/may-steinberg-project.pdf]  
 +
 
 +
:* Data Set of Question Classification [http://cogcomp.cs.illinois.edu/Data/QA/QC/],test the system of our QA
 +
:* read the paper "learning question classifiers "[http://www.aclweb.org/anthology/C/C02/C02-1150.pdf]
 +
 
 +
 
 +
==RNN Learn==
 +
* train the RNNLM using data of wsj_npdata
 +
 
 +
==ISCSLP 2014 PPT==
 +
* modify the figure
 +
* add the vsm result
 +
* add section of conclusion
 +
 
 +
==Summary==
 +
* something to be done
 +
# the method of vsm for document classification
 +
# the ppt figure problem

2014年9月2日 (二) 14:35的最后版本

QA system

  • Select the text question (100) and type the class label
  • i don't know how to label the clsses(hold),maybe the classes is not reasonable.
  • read the paper "Building a Question classifier for TREC-Stytle Question Answering System" [1]
  • Data Set of Question Classification [2],test the system of our QA
  • read the paper "learning question classifiers "[3]


RNN Learn

  • train the RNNLM using data of wsj_npdata

ISCSLP 2014 PPT

  • modify the figure
  • add the vsm result
  • add section of conclusion

Summary

  • something to be done
  1. the method of vsm for document classification
  2. the ppt figure problem