<?xml version="1.0" standalone="yes"?>
<Paper uid="J88-3002">
  <Title>MODELING THE USER IN NATURAL LANGUAGE SYSTEMS</Title>
  <Section position="2" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 INTRODUCTION
</SectionTitle>
    <Paragraph position="0"> Systems that use natural language as a means of communication must do so in a natural manner. One of the features of communication between people is that they acquire and use considerable knowledge about their conversational partners. In order for machines to interact with people in a comfortable, natural manner, they too will have to acquire and use knowledge of the people with whom they are interacting.</Paragraph>
    <Paragraph position="1"> Early research on natural language interfaces tended to view natural language as a &amp;quot;very high level&amp;quot; query language. One of the important results of research in the latter half of the 1970s (Waltz 1978, Kaplan 1982) is the realization that natural language communication is much more. The use of natural language for communication includes a host of conventions that must be followed in the dialog (Grice 1975). A person interacting with a computer via natural language will assume that these conventions are being followed, and will be quite unsatisfied if they are not. Most of these conventions require, in one way or another, that a conversational participant have particular knowledge about the goals, plans, capabilities, attitudes, and beliefs of the other person.</Paragraph>
    <Paragraph position="2"> This paper analyzes the role of user models in systems that interact with individual users in a natural language. Although the necessity of having and using a model of the user has been seen for some time, only within the last few years has it been actively pursued as a research topic. This research has been driven, in part, by attempts to create natural language interfaces to systems that can be characterized as cooperative problem solvers. Examples of such systems include intelligent interfaces to expert systems (Finin et al 1986, Carbonell et al 1983), database systems (Carberry 1985, Webber 1986), intelligent tutoring systems (Kass 1987b), and help and advisory systems (Wilensky et al 1984).</Paragraph>
    <Section position="1" start_page="0" end_page="0" type="sub_section">
      <SectionTitle>
1.1 AN OVERVIEW OF THIS PAPER
</SectionTitle>
      <Paragraph position="0"> In the remainder of this section, the kinds of user models and systems to be discussed in this paper will be characterized, including a general definition of a user model and an outline of how it can be used by a cooperative, interactive system that converses in natural language. The next section addresses the question &amp;quot;What is to be modeled?&amp;quot; by looking in some depth at the types of information that might be contained in a user model. These can be broadly classified as the user's goals (and the plans he may use to achieve them), capabilities, attitudes, and knowledge or belief. In section 3 a set of dimensions along which user models can Copyright 1988 by the Association for Computational Linguistics. Permission to copy without fee all or part of this material is granted provided that the copies are not made for direct commercial advantage and the CL reference and this copyright notice are included on the first page. To copy otherwise, or to republish, requires a fee and/or specific permission. 0362-613X/88/0100o-0503.00 Computational Linguistics, Volume 14, Number 3, September 1988 5 Kass and Finin Modeling the User in Natural Language Systems be classified is presented, while section 4 considers the methods that might be used to acquire information of the user, especially of his goals, plans, and beliefs.</Paragraph>
      <Paragraph position="1"> Section 5 considers several high-level features that have an impact on the design of a user modeling system, such as which participant in the interaction bears responsibility for ensuring the communication, or what the penalty for an error in the user model is. These considerations have an impact on the potential benefits and costs of employing a user model. The concluding section raises some issues that will require additional research in order to produce a powerful, general user modeling system.</Paragraph>
    </Section>
  </Section>
class="xml-element"></Paper>