<?xml version="1.0" standalone="yes"?>
<Paper uid="N06-1006">
  <Title>Learning to recognize features of valid textual entailments</Title>
  <Section position="3" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> During the last five years there has been a surge in work which aims to provide robust textual inference in arbitrary domains about which the system has no expertise. The best-known such work has occurred within the field of question answering (Pasca and Harabagiu, 2001; Moldovan et al., 2003); more recently, such work has continued with greater focus in addressing the PASCAL Recognizing Textual Entailment (RTE) Challenge (Dagan et al., 2005) and within the U.S. Government AQUAINT program.</Paragraph>
    <Paragraph position="1"> Substantive progress on this task is key to many text and natural language applications. If one could tell that Protestors chanted slogans opposing a free trade agreement was a match for people demonstrating against free trade, then one could offer a form of semantic search not available with current keyword-based search. Even greater benefits would flow to richer and more semantically complex NLP tasks.</Paragraph>
    <Paragraph position="2"> Because full, accurate, open-domain natural language understanding lies far beyond current capabilities, nearly all efforts in this area have sought to extract the maximum mileage from quite limited semantic representations. Some have used simple measures of semantic overlap, but the more interesting work has largely converged on a graphalignment approach, operating on semantic graphs derived from syntactic dependency parses, and using a locally-decomposable alignment score as a proxy for strength of entailment. (Below, we argue that even approaches relying on weighted abduction may be seen in this light.) In this paper, we highlight the fundamental semantic limitations of this type of approach, and advocate a multi-stage architecture that addresses these limitations. The three key limitations are an assumption of monotonicity, an assumption of locality, and a confounding of alignment and evaluation of entailment.</Paragraph>
    <Paragraph position="3"> We focus on the PASCAL RTE data, examples from which are shown in table 1. This data set contains pairs consisting of a short text followed by a one-sentence hypothesis. The goal is to say whether the hypothesis follows from the text and general background knowledge, according to the intuitions of an intelligent human reader. That is, the standard is not whether the hypothesis is logically entailed, but whether it can reasonably be inferred.</Paragraph>
  </Section>
class="xml-element"></Paper>