Multiple Instance Learning by Discriminative Training of Markov Networks
Abstract
We introduce a graphical framework for multiple instance learning (MIL) based on Markov networks. This framework can be used to model the traditional MIL definition as well as more general MIL definitions. Different levels of ambiguity -- the portion of positive instances in a bag -- can be explored in weakly supervised data. To train these models, we propose a discriminative max-margin learning algorithm leveraging efficient inference for cardinality-based cliques. The efficacy of the proposed framework is evaluated on a variety of data sets. Experimental results verify that encoding or learning the degree of ambiguity can improve classification performance.
Cite
Text
Hajimirsadeghi et al. "Multiple Instance Learning by Discriminative Training of Markov Networks." Conference on Uncertainty in Artificial Intelligence, 2013.Markdown
[Hajimirsadeghi et al. "Multiple Instance Learning by Discriminative Training of Markov Networks." Conference on Uncertainty in Artificial Intelligence, 2013.](https://mlanthology.org/uai/2013/hajimirsadeghi2013uai-multiple/)BibTeX
@inproceedings{hajimirsadeghi2013uai-multiple,
title = {{Multiple Instance Learning by Discriminative Training of Markov Networks}},
author = {Hajimirsadeghi, Hossein and Li, Jinling and Mori, Greg and Zaki, Mohamed H. and Sayed, Tarek},
booktitle = {Conference on Uncertainty in Artificial Intelligence},
year = {2013},
url = {https://mlanthology.org/uai/2013/hajimirsadeghi2013uai-multiple/}
}