Answer Selection for Questions with Multiple-Answer-Choices in Question Answering System Based on Textual Entailment Recognition

Abstract

Question Answering (QA) system is one of the most important and demanding tasks in the field of Natural Language Processing (NLP) that is concerned with answering questions posed in a natural language. In QA systems, the answer generation task generates a list of candidate answers to the user's question, in which only one answer is correct. Answer Selection is one of the main components of the QA, which is responsible for selecting the best answer choice from the candidate answers suggested by the system. However, the selection process can be very challenging. To address this challenge, we propose an approach to answer questions with multiple answer choices for QA systems based on Textual Entailment (TE) recognition. The approach consists of combining three feature sets to evaluate whether one of the candidate answers can be inferred from the text returned by the system. Our system was designed to utilize information on the lexical, syntactic and semantic level in order to recognize the entailment between the generated hypotheses (H) and the text (T). For performance evaluation of our method, a set of experiments has been conducted on the dataset provided by CLEF 2012 through the task of QA4MRE. The obtained results show that our method helps significantly to tackle the problem of Answer Selection in Question Answering system.