<?xml version="1.0"?>
<?xml-stylesheet type="text/css" href="http://www.cslt.org/mediawiki/skins/common/feed.css?303"?>
<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="zh-cn">
		<id>http://www.cslt.org/mediawiki/index.php?action=history&amp;feed=atom&amp;title=Chao_Xing_2016-01-25</id>
		<title>Chao Xing 2016-01-25 - 版本历史</title>
		<link rel="self" type="application/atom+xml" href="http://www.cslt.org/mediawiki/index.php?action=history&amp;feed=atom&amp;title=Chao_Xing_2016-01-25"/>
		<link rel="alternate" type="text/html" href="http://www.cslt.org/mediawiki/index.php?title=Chao_Xing_2016-01-25&amp;action=history"/>
		<updated>2026-04-03T23:52:34Z</updated>
		<subtitle>本wiki的该页面的版本历史</subtitle>
		<generator>MediaWiki 1.23.3</generator>

	<entry>
		<id>http://www.cslt.org/mediawiki/index.php?title=Chao_Xing_2016-01-25&amp;diff=18805&amp;oldid=prev</id>
		<title>Xingchao：以“Last Week: ---- Solution: * None. Plan to do:     1. Test DRAW in transfer large scale image style.     2. Code basic experiments in balance word frequency.     3. F...”为内容创建页面</title>
		<link rel="alternate" type="text/html" href="http://www.cslt.org/mediawiki/index.php?title=Chao_Xing_2016-01-25&amp;diff=18805&amp;oldid=prev"/>
				<updated>2016-01-25T00:48:09Z</updated>
		
		<summary type="html">&lt;p&gt;以“Last Week: ---- Solution: * None. Plan to do:     1. Test DRAW in transfer large scale image style.     2. Code basic experiments in balance word frequency.     3. F...”为内容创建页面&lt;/p&gt;
&lt;p&gt;&lt;b&gt;新页面&lt;/b&gt;&lt;/p&gt;&lt;div&gt;Last Week:&lt;br /&gt;
----&lt;br /&gt;
Solution:&lt;br /&gt;
* None.&lt;br /&gt;
Plan to do:&lt;br /&gt;
    1. Test DRAW in transfer large scale image style.&lt;br /&gt;
    2. Code basic experiments in balance word frequency.&lt;br /&gt;
    3. Find some unsupervised strategies in word sense vector.&lt;br /&gt;
    4. Test knowledge matrix factorization.&lt;br /&gt;
&lt;br /&gt;
This Week:&lt;br /&gt;
----&lt;br /&gt;
What have done:&lt;br /&gt;
    1. Test DRAW in transfer large scale image style.&lt;br /&gt;
       1. LSTM for reconstruction images.&lt;br /&gt;
       2. LSTM add random property :&lt;br /&gt;
             Results : normal distribution failure to reconstruct images.&lt;br /&gt;
                       prior distribution mean = 0, std = 0.1 reconstruct images but lost random property.&lt;br /&gt;
                       prior distribution mean = 0, std = 0.25 contain both reconstruction and random property. But the random property is just change a little edges of images.&lt;br /&gt;
                       Consider the cost function contain both reconstruction error and KL-Divergence between prior distribution and training distribution.&lt;br /&gt;
                       Lost = reconstruction + 0.25 * KL. The images shows clearly, and have some random property in edges.&lt;br /&gt;
                       Lost = reconstruction + 0.5 * KL. Some of images also shows clearly but others are fuzzy. &lt;br /&gt;
    2. Code basic experiments in balance word frequency.&lt;br /&gt;
    &lt;br /&gt;
Plan to do:&lt;br /&gt;
    1. Test DRAW in transfer large scale image style.&lt;br /&gt;
&lt;br /&gt;
    2. Prepare technique report in combine tf-idf vector as low frequency words and distributional vector as high frequency words in text classification and sentiment classification.&lt;/div&gt;</summary>
		<author><name>Xingchao</name></author>	</entry>

	</feed>