<?xml version="1.0" standalone="yes"?>
<Paper uid="W05-0203">
  <Title>A real-time multiple-choice question generation for language testing a preliminary study-</Title>
  <Section position="3" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> Multiple-choice question exams are widely used and are effective to assess students' knowledge, however it is costly to manually produce those questions.</Paragraph>
    <Paragraph position="1"> Naturally, this kind of task should be done with a help of computer.</Paragraph>
    <Paragraph position="2"> Nevertheless, there have been very few attempts to generate multiple-choice questions automatically. Mitkov et al.(2003) generated questions for a linguistics exam in a semi-automatic way and evaluated that it exceeds manually made ones in cost and is at least equivalent in quality. There are some other researches that involve generating questions with multiple alternatives (Dicheva and Dimitrova, 1998). But to the best of our knowledge, no attempt has been made to generate this kind of questions in a totally automatic way.</Paragraph>
    <Paragraph position="3"> This paper presents a novel approach to generate multiple-choice questions using machine learning techniques. The questions generated are those of fill-in-the-blank type, so it does not involve transforming declarative sentences into question sentences as in Mitkov's work. This simplicity makes the method to be language independent.</Paragraph>
    <Paragraph position="4"> Although this application can be very versatile, in that it can be used to test any kind of knowledge as in history exams, as a purpose of this research we limit ourselves to testing student's proficiency in a foreign language. One of the purposes of this research is to automatically extract important words or phrases in a text for a learner of the language.</Paragraph>
  </Section>
class="xml-element"></Paper>