<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0"><channel><description>Machine Learning-related surfings of Amir massoud Farahmand + other interesting things!</description><title>SoloGen</title><generator>Tumblr (3.0; @sologen)</generator><link>http://sologen.tumblr.com/</link><item><title>Advice for Graduate Students in Statistics</title><description>&lt;a href="http://thesilog.sologen.net/?p=221"&gt;Advice for Graduate Students in Statistics&lt;/a&gt;</description><link>http://sologen.tumblr.com/post/38080200</link><guid>http://sologen.tumblr.com/post/38080200</guid><pubDate>Wed, 11 Jun 2008 20:38:57 -0600</pubDate></item><item><title>Nonparametric Bayes Applications to Biostatistics</title><description>&lt;a href="http://thesilog.sologen.net/?p=219"&gt;Nonparametric Bayes Applications to Biostatistics&lt;/a&gt;</description><link>http://sologen.tumblr.com/post/37678267</link><guid>http://sologen.tumblr.com/post/37678267</guid><pubDate>Sun, 08 Jun 2008 21:38:59 -0600</pubDate></item><item><title>Reinforcement Learning blog</title><description>&lt;a href="http://thesilog.sologen.net/?p=217"&gt;Reinforcement Learning blog&lt;/a&gt;</description><link>http://sologen.tumblr.com/post/36600343</link><guid>http://sologen.tumblr.com/post/36600343</guid><pubDate>Fri, 30 May 2008 12:15:11 -0600</pubDate></item><item><title>Bracketing Entropy Bounds for Distribution Function</title><description>&lt;a href="http://thesilog.sologen.net/?p=210"&gt;Bracketing Entropy Bounds for Distribution Function&lt;/a&gt;</description><link>http://sologen.tumblr.com/post/35408593</link><guid>http://sologen.tumblr.com/post/35408593</guid><pubDate>Tue, 20 May 2008 02:33:52 -0600</pubDate></item><item><title>Embedding, Metric Entropy, etc.</title><description>&lt;a href="http://thesilog.sologen.net/?p=207"&gt;Embedding, Metric Entropy, etc.&lt;/a&gt;</description><link>http://sologen.tumblr.com/post/32929354</link><guid>http://sologen.tumblr.com/post/32929354</guid><pubDate>Sat, 26 Apr 2008 07:17:26 -0600</pubDate></item><item><title>Shannon Sampling and Learning Theory</title><description>&lt;a href="http://thesilog.sologen.net/?p=212"&gt;Shannon Sampling and Learning Theory&lt;/a&gt;</description><link>http://sologen.tumblr.com/post/32188470</link><guid>http://sologen.tumblr.com/post/32188470</guid><pubDate>Fri, 18 Apr 2008 15:59:49 -0600</pubDate></item><item><title>Statistical Performance of Support Vector Machines</title><description>&lt;a href="http://thesilog.sologen.net/?p=206"&gt;Statistical Performance of Support Vector Machines&lt;/a&gt;</description><link>http://sologen.tumblr.com/post/31376800</link><guid>http://sologen.tumblr.com/post/31376800</guid><pubDate>Thu, 10 Apr 2008 12:04:10 -0600</pubDate></item><item><title>Compression-related ideas in Machine Learning</title><description>&lt;a href="http://thesilog.sologen.net/?p=205"&gt;Compression-related ideas in Machine Learning&lt;/a&gt;</description><link>http://sologen.tumblr.com/post/30811001</link><guid>http://sologen.tumblr.com/post/30811001</guid><pubDate>Fri, 04 Apr 2008 14:55:01 -0600</pubDate></item><item><title>A few papers on estimation and control of robotic systems</title><description>&lt;a href="http://thesilog.sologen.net/?p=201"&gt;A few papers on estimation and control of robotic systems&lt;/a&gt;</description><link>http://sologen.tumblr.com/post/30531538</link><guid>http://sologen.tumblr.com/post/30531538</guid><pubDate>Tue, 01 Apr 2008 22:50:25 -0600</pubDate></item><item><title>Tumblr and I</title><description>&lt;a href="http://thesilog.sologen.net/?p=200"&gt;Tumblr and I&lt;/a&gt;</description><link>http://sologen.tumblr.com/post/30045924</link><guid>http://sologen.tumblr.com/post/30045924</guid><pubDate>Thu, 27 Mar 2008 21:48:02 -0600</pubDate></item><item><title>Approximate Dynamic Programming with Gaussian Processes</title><description>&lt;a href="http://thesilog.sologen.net/?p=199"&gt;Approximate Dynamic Programming with Gaussian Processes&lt;/a&gt;</description><link>http://sologen.tumblr.com/post/30045925</link><guid>http://sologen.tumblr.com/post/30045925</guid><pubDate>Thu, 27 Mar 2008 21:48:02 -0600</pubDate></item><item><title>Arthur C. Clarke</title><description>&lt;b&gt;“He never grew up; but he never stopped growing”&lt;/b&gt;We missed &lt;a href="http://en.wikipedia.org/wiki/Arthur_C._Clarke"&gt;Arthur C. Clarke&lt;/a&gt; a few hours ago. </description><link>http://sologen.tumblr.com/post/29258922</link><guid>http://sologen.tumblr.com/post/29258922</guid><pubDate>Wed, 19 Mar 2008 01:03:55 -0600</pubDate></item><item><title>Discrepancy, Chaining, and Subgaussian Processes</title><description>&lt;a href="http://wwwmaths.anu.edu.au/~mendelso/papers/discnew.pdf"&gt;Discrepancy, Chaining, and Subgaussian Processes&lt;/a&gt;: by S. Mendelson (2007). </description><link>http://sologen.tumblr.com/post/29161121</link><guid>http://sologen.tumblr.com/post/29161121</guid><pubDate>Tue, 18 Mar 2008 02:01:00 -0600</pubDate><category>Empirical Process</category><category>Chaining</category></item><item><title>Regularization in Kernel Learning</title><description>&lt;a href="http://wwwmaths.anu.edu.au/~mendelso/papers/MN29-02-08.pdf"&gt;Regularization in Kernel Learning&lt;/a&gt;: &lt;p&gt;a paper by S. Mendelson and J. Neeman (2008).&lt;/p&gt;
&lt;p&gt;I need to take a look at this paper because they stated their results when the regularization term is not the usual RKHS norm squared. I’m wondering if their results can exploit “sparsity”.&lt;/p&gt;</description><link>http://sologen.tumblr.com/post/29159736</link><guid>http://sologen.tumblr.com/post/29159736</guid><pubDate>Tue, 18 Mar 2008 01:35:00 -0600</pubDate><category>RKHS</category><category>Kernel Methods</category><category>Convergence Rate</category></item><item><title>On Cross-Validation</title><description>Take a look at &lt;a href="http://hunch.net/index.php?p=29"&gt;these papers&lt;/a&gt; (from &lt;a href="http://hunch.net/"&gt;John Langford’s blog&lt;/a&gt;).</description><link>http://sologen.tumblr.com/post/28962182</link><guid>http://sologen.tumblr.com/post/28962182</guid><pubDate>Sat, 15 Mar 2008 19:12:06 -0600</pubDate><category>Cross-Validation</category></item><item><title>Rademacher Complexities and related stuff</title><description>&lt;p&gt;A few sources that talk about Rademacher complexities (and other stuff):&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt; O. Bousquet, S. Boucheron, and G. Lugosi, “&lt;a href="http://www.kyb.mpg.de/publications/pdfs/pdf2819.pdf"&gt;Introduction to Statistical Learning Theory,&lt;/a&gt;” .&lt;/li&gt;
&lt;li&gt;P. Bartlett, O. Bousquet, and S. Mendelson, “&lt;a href="http://arxiv.org/pdf/math.ST/0508275"&gt;Local Rademacher Complexities,” The Annals of Statistics&lt;/a&gt;,” 2005.&lt;/li&gt;
&lt;li&gt;V. Koltchinskii and D. Panchenko, “&lt;a href="http://arxiv.org/pdf/math/0405338v1"&gt;Rademacher Processes and Bounding the Risk of Function Learning,&lt;/a&gt;” ?.&lt;/li&gt;
&lt;li&gt;P. Bartlett and S. Mendelson, “&lt;a href="http://www.ai.mit.edu/projects/jmlr/papers/volume3/bartlett02a/bartlett02a.pdf"&gt;Rademacher and Gaussian Complexities: Risk Bounds and Structural Results,&lt;/a&gt;” JMLR 2002 (I need to check this paper, especially because it defines Gaussian complexity. Is there any generalization for arbitrary “noise” term (instead of Gaussian or boolean)? )&lt;/li&gt;
&lt;/ul&gt; </description><link>http://sologen.tumblr.com/post/28606553</link><guid>http://sologen.tumblr.com/post/28606553</guid><pubDate>Tue, 11 Mar 2008 20:52:00 -0600</pubDate><category>Concentration Inequalities</category><category>Empirical Processes</category><category>Learning Theory</category></item><item><title>Concentration Inequalities and Model Selection</title><description>&lt;a href="http://www.math.u-psud.fr/~massart/stf2003_massart.pdf"&gt;Concentration Inequalities and Model Selection&lt;/a&gt;: A book by &lt;a href="http://www.math.u-psud.fr/~massart/"&gt;Pascal Massart&lt;/a&gt;.</description><link>http://sologen.tumblr.com/post/28600349</link><guid>http://sologen.tumblr.com/post/28600349</guid><pubDate>Tue, 11 Mar 2008 19:03:37 -0600</pubDate></item><item><title>On the L_1-L_q Regularized Regression</title><description>&lt;a href="http://www.stat.cmu.edu/tr/tr860/tr860.pdf"&gt;On the L_1-L_q Regularized Regression&lt;/a&gt;: a technical report by Han Liu and Jian Zhang (2008).</description><link>http://sologen.tumblr.com/post/28510149</link><guid>http://sologen.tumblr.com/post/28510149</guid><pubDate>Mon, 10 Mar 2008 21:04:29 -0600</pubDate></item><item><title>Life Goes on in Tehran</title><description>&lt;a href="http://www.lifegoesonintehran.com/index.html"&gt;Life Goes on in Tehran&lt;/a&gt;: &lt;p&gt;A photoblog about Tehran and Iran.&lt;/p&gt;
&lt;p&gt;Yes! It is not immediately related to machine learning. (;&lt;/p&gt;</description><link>http://sologen.tumblr.com/post/28036613</link><guid>http://sologen.tumblr.com/post/28036613</guid><pubDate>Wed, 05 Mar 2008 10:58:12 -0700</pubDate><category>Iran</category><category>Photo</category></item><item><title>Some References for Cross-Validation</title><description>&lt;ul&gt;
&lt;li&gt;B. Efron and G. Gong, “&lt;a href="http://www.jstor.org/view/00031305/di020575/02p0141f/0"&gt;A Leisurely Look at the Bootstrap, the Jackkife, and Cross-Validation,&lt;/a&gt;” 1983.&lt;/li&gt;
&lt;li&gt;M. Kearns, “&lt;a href="http://www.cis.upenn.edu/~mkearns/papers/cv.pdf"&gt;A Bound on the Error of Cross Validation Using the Approximation and Estimation Rates, with the Consequences for the Training-Test Split,&lt;/a&gt;” 1997.&lt;/li&gt;
&lt;li&gt;P. Burman, E. Chow, and D. Nolan, “&lt;a href="http://biomet.oxfordjournals.org/cgi/reprint/81/2/351"&gt;A Cross-Validatory Method for Dependent Data,&lt;/a&gt;” 1994&lt;/li&gt;
&lt;li&gt;B. Efron and R. Tibshirani, “&lt;a href="http://www.jstor.org/view/01621459/di986012/98p0145k/0"&gt;Improvements on Cross-Validation: The 0.632+ Bootstrap Method&lt;/a&gt;,” 1997.&lt;/li&gt;
&lt;li&gt;G. Golub, M. Heath, and G. Wahba, “&lt;a href="http://www.jstor.org/view/00401706/ap040083/04a00090/0"&gt;Generalized Cross-Validation as a Method for Choosing a Good Ridge Regression,&lt;/a&gt;” 1979. &lt;/li&gt;
&lt;li&gt;J. Shao, “&lt;a href="http://www.jstor.org/view/01621459/di985996/98p0310q/0"&gt;Linear Model Selection by Cross-Validation,&lt;/a&gt;” 1993.&lt;/li&gt;Leave-one-out CV is not consistent! He suggests to use n_v/n —&gt; 1 method while n —&gt; \infty.&lt;/ul&gt;</description><link>http://sologen.tumblr.com/post/27897194</link><guid>http://sologen.tumblr.com/post/27897194</guid><pubDate>Mon, 03 Mar 2008 21:23:00 -0700</pubDate></item></channel></rss>
