HTTPS SSH
This code implements the submodular function maximization using modified greedy algorithm.
The code is used for various subset selection tasks, such as data selection, and feature/ attribute selection.

Features:
* Subset selection via submodular function maximization
* The functions include: facility location function, saturate coverage function, softmax function, and an optional diversity reward function
* Support submodular function maximization with knapsack or cardinality constraints
* Speed up with accelerated greedy algorithm (Minoux 1976)

Requirement:
    C++11 or Boost library

Install:
	 make all;

Written by Yuzong Liu
Based on Hui Lin's implementation


- If you use the code for your research, we recommend you cite the following papers:

@inproceedings{liu2013submodular,
  title={Submodular Feature Selection for High-Dimensional Acoustic Score Space},
  author={Liu, Yuzong and Wei, Kai and Kirchhoff, Katrin and Song, Yisong and Bilmes, Jeff},
  booktitle={Proceedings of IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP)},
  year={2013}
}

@inproceedings{kirchhoff2013classification,
  title={Classification of developmental disorders from speech signals using submodular feature selection.},
  author={Kirchhoff, Katrin and Liu, Yuzong and Bilmes, Jeff},
  booktitle={Proceedings of Annual Conference of the International Speech Communication Association (Interspeech)},
  year={2013}
}

@inproceedings{dsss2013,
  title={Using Document Summarization Techniques for Speech Data Subset Selection.},
  author={Wei, Kai and Liu, Yuzong and Kirchhoff, Katrin and Bilmes, Jeff},
  booktitle={Proceedings of NAACL-HLT},
  year={2013}
}