1669 search results - page 296 / 334 » SMO-Style Algorithms for Learning Using Privileged Informati... |
KDD
2012
ACM
11 years 10 months ago
2012
ACM
Multi-instance multi-label learning (MIML) is a framework for supervised classification where the objects to be classified are bags of instances associated with multiple labels....
CIKM
2011
Springer
12 years 7 months ago
2011
Springer
Traditional feature selection methods assume that the data are independent and identically distributed (i.i.d.). In real world, tremendous amounts of data are distributed in a net...
IJON
1998
13 years 7 months ago
1998
In this paper we employ information theoretic algorithms, previously used for separating instantaneous mixtures of sources, for separating convolved mixtures in the frequency doma...
ECCV
2008
Springer
14 years 9 months ago
2008
Springer
Recently, models based on conditional random fields (CRF) have produced promising results on labeling sequential data in several scientific fields. However, in the vision task of c...
ICDM
2009
IEEE
14 years 2 months ago
2009
IEEE
Abstract -- Detection of execution anomalies is very important for the maintenance, development, and performance refinement of large scale distributed systems. Execution anomalies ...