This shows you the differences between two versions of the page.

groupnews:seminar_on_local_global_and_hybrid_learning_at_institute_of_information_science_academia_sinica_taiwan_on_november_4_2005 [2005/11/05 00:00] (current)
Line 1: Line 1:
 +====== Seminar on "Local, Global, and Hybrid Learning" at Institute of Information Science, Academia Sinica, Taiwan on November 4, 2005... ======
 +The use of classifiers permeates in various fields of
 +engineering and science disciplines.  When constructing
 +classifiers, there is a dichotomy in choosing whether to
 +use local vs. global characteristics of the input data.  In
 +this talk, we will describe our work on combining the local
 +and global learning in the Maxi-Min Margin Machine (M^4).  
 +M^4 presents a unifying theory that subsumes the Support
 +Vector Machine (SVM), the Minimax Probability Machine
 +(MPM), and the Linear Discriminant Analysis (LDA).  While
 +LDA and MPM focus on building the decision plane using
 +global information and SVM focuses on building the decision
 +plane in a local manner, M^4 incorporates these two
 +seemingly different yet complementary characteristics in a
 +unifying framework that achieves good classification
 +accuracy.  We will present the formulation of M^4 and also
 +experimental results to show the advantage of our novel
 +  * [[http://www.iis.sinica.edu.tw/Function/seminar-new.php?TYPE=SE]]
 +  * [[http://www.iis.sinica.edu.tw/HTML/ENGLISH/S200511041000.html]]
groupnews/seminar_on_local_global_and_hybrid_learning_at_institute_of_information_science_academia_sinica_taiwan_on_november_4_2005.txt · Last modified: 2005/11/05 00:00 (external edit)     Back to top