en
0.25
0.5
0.75
1.25
1.5
1.75
2
Compact Coding for Hyperplane Classifiers in Heterogeneous Environment
Published on Oct 03, 20113085 Views
Transfer learning techniques have witnessed a significant development in real applications where the knowledge from previous tasks are required to reduce the high cost of inquiring the labeled informa
Related categories
Chapter list
Compact Coding for Hyperplane Classifiers in Heterogeneous Environment00:00
Inductive Transfer Learning with multiple source tasks (1)00:18
Inductive Transfer Learning with multiple source tasks (2)01:23
Problems of the Negative Transfer (NT) - 101:43
Problems of the Negative Transfer (NT) - 202:33
Existing Methods and the Objective of our algorithm (1)03:06
Existing Methods and the Objective of our algorithm (2)03:33
Problem Setting and our Motivation03:56
A Simple Example04:35
Minimum Description Length Principle (MDLP) [Quinlan 89]05:11
Compact Coding for Hyperplane Classifiers (CCHC)06:19
Code Length as the Similarity Measure (1)07:05
Code Length as the Similarity Measure (2)07:23
Code Length as the Similarity Measure (3)07:34
Code Length as the Similarity Measure (4)07:39
Code Length as the Similarity Measure (5)08:10
Code Length as the Similarity Measure (6)08:19
Preliminaries of coding08:34
Coding method of CCHC09:25
Calculation of the code length of the toy example09:55
Algorithm CCHC10:17
Experimental setting11:02
Results on mushroom data sets11:46
Results of kr vs kp and splice12:38
Results for rec vs talk as the target data set13:00
Results for rec vs sci, and talk vs sci as the target data set13:23
Transferred components in text data sets in Micro Level13:34
Summary of this work14:16
Thank you!14:52