<?xml version="1.0" standalone="yes"?>
<Paper uid="W00-0311">
  <Title>A Compact Architecture for Dialogue Management Based on Scripts and Meta-Outputs</Title>
  <Section position="2" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 Introduction 1
</SectionTitle>
    <Paragraph position="0"> The basic task we consider in this paper is that of using spoken language to give commands to a semi-autonomous robot or other similar system. As evidence of the importance of this ta~k in the NLP community note that the early, influential system SHRDLU (Winograd, 1973) was intended to address just this type of problem. More recent work on spoken language interfaces to semi-autonomous robots include SRI's Flakey robot (Konolige et al., 1993) and NCARAI's InterBOT project (Perzanowski et al., 1998; Perzanowski et al., 1999). A number of other systems have addressed part of the task. CommandTalk (Moore et al., 1997), Circuit Fix-It Shop (Smith, 1997) and Tl:tAINS-96 (Traum and Allen, 1994; Traum and Andersen, 1999) are spoken language systems but they interface to simulation or help facilities rather than semi-autonomous agents.</Paragraph>
    <Paragraph position="1"> Jack's MOOse Lodge (Badler et al., 1999) takes text rather than speech as natural language input and the avatars being controlled are not semi-autonomous.</Paragraph>
    <Paragraph position="2"> Other researchers have considered particular aspects of the problem such as accounting for various aspects of actions (Webber, 1995; Pym et al., 1995). In most of this and other related work the treatment is some variant of the following. If there is a speech interface, the input speech signal is converted into text. Text either from the recognizer or directly input by IThis paper also appears in the proceedings of the Sixth International Conference on Applied Natural Language Processing~ Seattle, WA, April 2000.</Paragraph>
    <Paragraph position="3"> the user is then converted into some kind of logical formula, which abstractly represents the user's intended command; this formula is.then fed into a command interpreter, which execdtes the command.</Paragraph>
    <Paragraph position="4"> We do not think the standard treatment outlined above is in essence incorrect, but we do believe that, as it stands, it is in-need of some modification. This paper will in particular make three points. First, we suggest that the output representation should not be regarded as a logical expression, but: rather as a program in some kind of scripting language. Second, we argue that it is not merely the case that the process of converting the input signal to the final representation can sometimes go wrong; rather, this is the normal course of events, and the inferpretatiofi process should be organized with that assumption in mind. Third, we claim, perhaps surprisingly, that the first and second points are related. These claims are elaborated in Section 2.</Paragraph>
    <Paragraph position="5"> The remainder of the paper describes an architecture which addresses the issues outlined above, and which has been used to implement a prototype speech interface to a simulated semi-autonomous robot intended for deployment on the International Space Station. Sections 3 and 4 present an overview of the implemented interface, focussingon representational issues relevant to dialogue management. Illustrative examples of interactions with the system are provided in Section 5. Section 6 concludes.</Paragraph>
  </Section>
class="xml-element"></Paper>