It is a pleasure to announce that a short paper describing KeLP has been published in the Journal of Machine Learning Research. You can download the paper at the following link.

A new version of KeLP has been released! This release includes: A new kernel cache that is meant to substitute the FixIndexKernelCache and the DynamicIndexKernelCache. New Evaluators for measuring the quality of Clustering algorithms. Added a new way of serializing/desearlizing Json files in gzip format. As usual, KeLP packages can be found in our Maven Read more about KeLP 2.2.1 released[…]

If you need support to generate input material that involves tree structure, you should take a look at this page.    

A system completely developed in KeLP participated to the Community Question Answering task at SemEval 2017. It achieved the first position in subtask A (i.e., answer selection), first position in subtask B (i.e., finding matching questions) and third position in subtask C (i.e., finding previously posted comments answering to a new question). The complete results are here.

A new version of KeLP has been released! In addition to minor bug fixes, this release includes: A new learning algorithm that enable (for the first time in KeLP) to deal with sequences labeling problems! It is based on a Markovian formulation within a SVM framework. Most noticeably: the new algorithm can deal both with linear Read more about KeLP 2.2.0 released[…]