SemiL

Semi-supervised Learning Software:

The free semi-supervised learning or transductive inference software SemiL (written by Te Ming Huang and developed by Te Ming Huang and Vojislav Kecman, with a support of Dr. Chan-Kyoo Park. Discussion with and help of Dr. Dengyong Zhou is highly appreciated) implements, extends and improves two approaches presented in papers,

Zhou, D., Bousquet, O., Lal, T. N., Weston, J., Schölkopf, B.:
Learning with Local and Global Consistency, NIPS 16, pp. 321-328, 2004

Zhu, X.-J., Ghahramani, Z., Lafferty, J.:
Semi-Supervised Learning Using Gaussian Fields and Harmonic Functions, ICML 2003

and it was initiated during the short stay of the first author at the Max Planck Institute (Department B. Schölkopf).

SemiL is efficient software for solving large scale semi-supervised learning or transductive inference problems using graph based approaches when faced with unlabeled data. It implements various semi-supervised learning approaches as listed below:

Hard label approach with the maximization of smoothness, and
Soft label approach with the maximization of smoothness,
for all three types of models (i.e., Basic model, Norm Constrained Model and Bound Constrained Model) by using either Standard or Normalized Laplacian.

Download the zipped software (SemiL.zip 1.7MB) by clicking here.
After unzipping, it will self-extract itself into the folder SemiL. Read the User Manual first, and play with SemiL under Windows (MS-DOS, Command Prompt) or under a Linux platform.

If you would like to cite SemiL, please cite the following document:

Huang T.-M., Kecman V., SemiL, Semi-supervised Learning Software, Manual, http://www.learning-from-data.com/te-ming/semil.htm, 2005.

back home