Using Recognizing Textual Entailment as a Core Engine for Answer Validation

Abstract

Abstract. This paper is about our approach to answer validation, which centered by a Recognizing Textual Entailment (RTE) core engine. We first combined the question and the answer into Hypothesis (H) and view the document as Text (T); then, we used our RTE system to check whether the entailment relation holds between them. Our system was evaluated on the Answer Validation Exercise (AVE) task and achieved f-measures of 0.46 and 0.55 for two submission runs, which both outperformed others' results for the English language.

JAIST: Combining multiple features for Answer Selection in Community Question Answering

“…The task of identifying answer quality has been studied by many researchers in the field of Question Answering. Many methods have been proposed: web redundancy information (Magnini et al, 2002), non-textual features (Jeon et al, 2006), textual entailment (Wang and Neumann, 2007), syntactic features (Grundström and Nugues, 2014). However, most of these works used independent dataset and evaluation metrics; thus it is difficult to compare the results of these methods.…”

GPT-4 passes the bar exam

“…Building up the broader concept of textual entailment [74][75][76], earlier work studying Bar Exams treated 'the relationship between the question and the multiple-choice answers as a form of textual entailment' [77] where the ability to identify wrong answers (non-entailment) is differentiated from the ability to identify the correct answer (entailment). Intuitively, this is related to the classic test taking strategy of eliminating clearly erroneous answers.…”