78 search results - page 6 / 16 » Multiple instance learning for sparse positive bags |
NPL
2006
13 years 7 months ago
2006
In multi-instance learning, the training examples are bags composed of instances without labels, and the task is to predict the labels of unseen bags through analyzing the training...
ICML
2008
IEEE
14 years 8 months ago
2008
IEEE
In multiple instance learning (MIL), how the instances determine the bag-labels is an essential issue, both algorithmically and intrinsically. In this paper, we show that the mech...
MIR
2004
ACM
14 years 1 months ago
2004
ACM
In this paper, we propose an autonomous learning scheme to automatically build visual semantic concept models from the output data of Internet search engines without any manual la...
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....
AAAI
2008
13 years 10 months ago
2008
Multiple instance learning (MIL) is a branch of machine learning that attempts to learn information from bags of instances. Many real-world applications such as localized content-...