Dear all, We are delighted to announce our recently released ITE (Information Theoretical Estimators) toolbox.
The ITE package could be of interest to many of you: the estimation of information theoretical quantities (entropy, mutual information, divergence) plays a central role in numerous important problems of machine learning. Unfortunately, available packages focus on (i) discrete variables, or (ii) quite specialized applications and information theoretical estimation methods. To fill in this serious gap, we have recently released ITE (i) a highly modular, (ii) free and open source, (iii) multi-platform toolbox, which 1. is capable of estimating many different variants of entropy, mutual information and divergence measures. 2. offers a simple and unified framework to (a) easily construct new estimators from existing ones or from scratch, and (b) transparently use the obtained estimators in information theoretical optimization problems. 3. with a prototype application in a central problem family of signal processing, independent subspace analysis and its extensions. The homepage of ITE is "https://bitbucket.org/szzoli/ite/";. Feel free to use it. Best, Zoltan ("http://nipg.inf.elte.hu/szzoli";) _______________________________________________ uai mailing list [email protected] https://secure.engr.oregonstate.edu/mailman/listinfo/uai
