A Surface-Similarity Based Two-Step Classifier for RITE-VAL

This paper describes the system of the team SKL in the NTCIR-11 RITE-VAL workshop. The system consists of two modules: RTE module and text-search module. The RTE module, which is a modified version of our previous system for the binary classification in the RITE-2 workshop, takes two-step classification strategy. The first step classifies a given text pair into positive or negative entailment class based on an overlap measure. If the pair is classified into positive class, the second step examines whether the assigned class should be flipped or not by using heuristic rules that detect the mismatch of named entities and numbers. The Fact Validation subtask in this workshop is to determine whether a given hypothesis is true or not based on a given document. For this subtask, we introduce the text-search module that extracts the text segment from the document; the RTE module produces the final output from the extracted text segment and the hypothesis.