<?xml version="1.0" standalone="yes"?>
<Paper uid="J82-3002">
  <Title>An Efficient Easily Adaptable System for Interpreting Natural Language Queries 1</Title>
  <Section position="6" start_page="0" end_page="0" type="evalu">
    <SectionTitle>
5. Performance and Portability
</SectionTitle>
    <Paragraph position="0"> Chat-80 at its current stage of implementation, covers a limited but useful subset of English. A fair idea of the range of the present subset is given by the examples in Appendix I. It will be seen that the sub-set includes nouns, verbs, adjectives, prepositions, and determiners, with a fairly full coverage of interrogative and relative constructions. We have concentrated on features of English that seemed essential for simple question answering; there are many directions in which the subset could usefully be extended and which do not appear to pose any particular difficulties.</Paragraph>
    <Paragraph position="1"> At present, the system accepts a small vocabulary of about 100 domain dependent works (not counting proper nouns, but including alternative word forms such as plurals). This vocabulary can very easily be extended (as indicated below). In addition there are some 50 domain independent words. On the whole, any question that can be expressed using this vocabulary is correctly understood and answered by the system. null The sizes of the different components of the system are indicated below in terms of the approximate num- null The speed of Chat-80 on some sample queries relating to the geographical database is shown in Appendix I. Generally speaking, any query in this domain that can comfortably be expressed in a single sentence of the English subset is answered in well under one second of CPU time. Note that the domain dependent vocabulary could be much extended without having any significant impact on these times (because of the way the dictionary is indexed).</Paragraph>
    <Paragraph position="2"> It is also worth noting that, for all but the simplest queries, natural language analysis represents only a small proportion of the total time. This suggests that, as far as the efficiency of natural language question answering systems is concerned, it is the answering process rather than the natural language analysis to which most effort needs to be directed. Certainly this has been our approach, although it appears to be somewhat contrary to the prevailing view in artificial intelligence. In particular, parsing does not seem to pose any major efficiency problem, provided one does not expect the grammar to do too much.</Paragraph>
    <Paragraph position="3"> As regards portability, we think Chat-80 should be relatively easy to adapt to different applications - for the same reasons that we found it easy to adapt Dahl's program to English and to a different domain.</Paragraph>
    <Paragraph position="4"> Partly this is due to the fact that (in both systems) the domain dependent parts are clearly separated from the rest of the system, and are broken down into small units which can be added incrementally as &amp;quot;data&amp;quot; (see Appendix II). Thus our natural language analysis modules deal exclusively with general features of English, in contrast to the &amp;quot;semantic grammar&amp;quot; approach (Burton 1976).</Paragraph>
    <Paragraph position="5"> Now there are other practical systems which have not taken the &amp;quot;semantic grammar&amp;quot; approach but are, we feel, less easy to modify than Chat-80; LUNAR (Woods, Kaplan, and Nash-Webber 1972) is a good example. The reason lies in the way &amp;quot;meanings&amp;quot; are attached to words. In LUNAR, a meaning is simply a procedure. For nouns it is a procedure to generate objects in a certain class; for most other words it is a procedure to test whether some property is true of given objects. This entails that &amp;quot;meanings&amp;quot; can be executed in only one way, and precludes the kind of query planning done in Chat-80. But such a simple-minded approach to query execution is not viable in most practical situations, as Woods 1977 recognises with his &amp;quot;smart quantifiers&amp;quot;. The only alternative, given the procedural approach to meaning, is to represent the meaning of a word by a set of alternative procedures to be used in different circumstances. But this makes life very difficult for someone wanting to introduce a new word or concept into the system.</Paragraph>
    <Paragraph position="6"> In Chat-80, the meaning of a new word is in principle just a set of facts and general rules that define the predicate corresponding to that word. The procedural aspect is on the whole taken care of by the planning algorithm, and by Prolog's flexible handling of predications in which only certain arguments are instantiated. However, with the present system, the definer of a new word must be responsible for ensuring that the predicate definition he supplies is not only correct, but is also reasonably efficient when executed by Prolog.</Paragraph>
  </Section>
class="xml-element"></Paper>