kernelized methods in natural language processing

Question

I am new to NLP. I want to implement a matching approach for short sentences( e.g. questions in cQA). I want to use tree kernel function as a syntactic feature. I am wondering to know is there any implementation available in NLP tools, or elsewhere?

Specifically, I like a method like this paper Collins and Duffy, Convolution kernels for natural language processing.

Any suggestion is useful and appreciated.

Answer 1

It is common to compare two tree structures to find similarity between sentences in natural language processing. A good tutorial has been written by Alessandro Moschitti at SVM-LIGHT-TK . Kernel methods have been widely used in machine learning tasks ( e.g. SVM ), and thus there are different approach of kernelization adapted for various tasks.