<?xml version="1.0" standalone="yes"?>
<Paper uid="H05-1009">
  <Title>Proceedings of Human Language Technology Conference and Conference on Empirical Methods in Natural Language Processing (HLT/EMNLP), pages 65-72, Vancouver, October 2005. c(c)2005 Association for Computational Linguistics NeurAlign: Combining Word Alignments Using Neural Networks</Title>
  <Section position="10" start_page="71" end_page="71" type="concl">
    <SectionTitle>
6 Conclusions
</SectionTitle>
    <Paragraph position="0"> We presented NeurAlign, a novel approach to combining the outputs of different word alignment systems. Our approach treats individual alignment systems as black boxes, and transforms the individual alignments into a set of data with features that are borrowed from their outputs and additional linguistic features (such as POS tags and dependency relations). We use neural nets to learn the true alignments from these transformed data.</Paragraph>
    <Paragraph position="1"> We show that using POS tags to partition the transformed data, and learning a different classifier for each partition is more effective than using the entire data at once. Our results indicate that NeurAlign yields a significant 28-39% relative error reduction over the best of the input alignment systems and a significant 20-34% relative error reduction over the best known alignment combination technique on English-Spanish and English-Chinese data.</Paragraph>
    <Paragraph position="2"> We should note that NeurAlign is not a stand-alone word alignment system but a supervised learning approach to improve already existing alignment systems. A drawback of our approach is that it requires annotated data. However, our experiments have shown that significant improvements can be obtained using a small set of annotated data. We will do additional experiments to observe the effects of varying the size of the annotated data while learning neural nets. We are also planning to investigate whether NeurAlign helps when the individual aligners are trained using more data.</Paragraph>
    <Paragraph position="3"> We will extend our combination approach to combine word alignment systems based on different models, and investigate the effectiveness of our technique on other language pairs. We also intend to evaluate the effectiveness of our improved alignment approach in the context of machine translation and cross-language projection of resources.</Paragraph>
    <Paragraph position="4"> Acknowledgments This work has been supported in part by ONR MURI Contract FCPO.810548265, Cooperative Agreement DAAD190320020, and NSF ITR Grant IIS0326553. null</Paragraph>
  </Section>
class="xml-element"></Paper>