<?xml version="1.0" standalone="yes"?>
<Paper uid="W99-0405">
  <Title>Modeling the language assessment process and result: Proposed architecture for automatic oral proficiency assessment</Title>
  <Section position="9" start_page="29" end_page="29" type="concl">
    <SectionTitle>
7 Conclusions
</SectionTitle>
    <Paragraph position="0"> We have outlined challenges for modeling the human rating task, both in terms of process and result. In the domain of acoustic features and speech recognition, we suggest the technology currently does not permit complete modeling of the rating process. Nevertheless, the paucity of data in our domain requires us to adopt the transfer model of speech, which permits automatic adaptation to errorful speech.</Paragraph>
    <Paragraph position="1"> In addition, our recognizer incorporates a relaxed grammar, permitting input to vary from the target language at the lexical and syntactic levels. These adaptations allow us to model human perception and processing of (non-native) speech, as required by our task. In the non-acoustic domain, we also adopt machine learning techniques to pool the relevant humanidentified features. As a result, we can learn more about the feature-weighting in the process of tuning to an appropriate level of reliability with the results of human raters.</Paragraph>
  </Section>
class="xml-element"></Paper>