This the end day of 2003, tomorrow is a new year.
New year, will be new vision.
I'll try again!
2003年12月30日
First Volume of FSNLP
Last midnight, I finished the reading of the first volume of FSNLP.
This morning, I changed it for the second volume.
The first book gives me a summary of FSNLP. And I have found out that FSNLP is very good for the primary researcher to read for being familiar with NLP.
With the Adding Knowledge Program, I have persisted and gained a little.
I will keep on.
This morning, I changed it for the second volume.
The first book gives me a summary of FSNLP. And I have found out that FSNLP is very good for the primary researcher to read for being familiar with NLP.
With the Adding Knowledge Program, I have persisted and gained a little.
I will keep on.
2003年12月29日
YOCSEF Meeting
This morning, I arrived at the administration building's YOCSEF assembly room.
After we had disposed the assembly room, the meeting's chairman Pro. Zhao Tiejun declare the begin of YOCSEF. This is my second time to attend YOCSEF.
The subject of this YOCSEF is Digital Olympics and muitillanguage information processing. And there were three specially invited guests. They described the blueprint of Olympics and the current most difficulties.
I have been familiar with the general picture of the Digital Olympics. And I found out that the Information Processing is more useful in the current times. And the prospect of our lab is very beautiful.
After we had disposed the assembly room, the meeting's chairman Pro. Zhao Tiejun declare the begin of YOCSEF. This is my second time to attend YOCSEF.
The subject of this YOCSEF is Digital Olympics and muitillanguage information processing. And there were three specially invited guests. They described the blueprint of Olympics and the current most difficulties.
I have been familiar with the general picture of the Digital Olympics. And I found out that the Information Processing is more useful in the current times. And the prospect of our lab is very beautiful.
2003年12月28日
2003年12月27日
2003年12月26日
Rough Set Theory
Our WSD research group will do some experiments on Rough Set. And Mr.Lzm believes that rough set can be used for WSD very effectively.
So I am arranged to read dome materials on RS.
Firstly, I read Knowledge Discover by TSinghua Express. But the content of this book is not enough. So I find some papers about RS.
I have understood that reading summarize article is the fast way to be familiar with the area. I read a summarize paper about RS. And the knowledge points are very clearly.
So I am arranged to read dome materials on RS.
Firstly, I read Knowledge Discover by TSinghua Express. But the content of this book is not enough. So I find some papers about RS.
I have understood that reading summarize article is the fast way to be familiar with the area. I read a summarize paper about RS. And the knowledge points are very clearly.
2003年12月25日
Studying paper of Li Juanzi
This morning I am reading Li Juanzi's paper 《语言模型中的一种改进的最大熵方法及其应用》which is published on JOURNAL OF SOFTWARE.
In her paper, she used an updated method combining maximum entropy, mutual information and Z-test to choose the best feature of context for a multivocal word and then used IIS algorithm to optimize the parameters of the linear model for Word Sense Disambiguation.
The experiment results are displaying the advantage of this approach. But I think the paper has two flaws. Firstly, the experiments for WSD is not enough. Secondly, Z-test is usually used to test normal distribution for large scale's samples. And in this paper there is a connotative hypothesis that the mutual information between the feature set and the category a multivocal word is followed normal distribution. The experiments did not prove this hypothesis.
And I think this experiments could be done more fully, and the experiment should do the hypothesis test.
OK. There is an idea. We can arrange some little experiments of our current information corpus to do some hypothesis test. En, this idea should be thought more and more.
In her paper, she used an updated method combining maximum entropy, mutual information and Z-test to choose the best feature of context for a multivocal word and then used IIS algorithm to optimize the parameters of the linear model for Word Sense Disambiguation.
The experiment results are displaying the advantage of this approach. But I think the paper has two flaws. Firstly, the experiments for WSD is not enough. Secondly, Z-test is usually used to test normal distribution for large scale's samples. And in this paper there is a connotative hypothesis that the mutual information between the feature set and the category a multivocal word is followed normal distribution. The experiments did not prove this hypothesis.
And I think this experiments could be done more fully, and the experiment should do the hypothesis test.
OK. There is an idea. We can arrange some little experiments of our current information corpus to do some hypothesis test. En, this idea should be thought more and more.
订阅：
博文 (Atom)