Framework independent summarized parser output in XML and its example-based documentation

Tam Wai Lok, Yusuke Miyao, and Juniichi Tsujii

Abstract

We see a communication problem between the grammar engineering community and the NLP community. The information to be communicated is the results produced by a grammar. This paper is about our solution to the problem. Our solution has two components: an alternative output format and its documentation. Our alternative output format carries constituency information that parsers are built for computing, but in lesser quantity and a simpler form than the standard AVM output format. The documentation for it provides shallow and static explanation different from the deep and dynamic explanation found in literatures about grammar formalisms meant for grammar writers. The shallow and static explanation is meant to enable members of the NLP community to achieve a shallow level of understanding of the results produced by a grammar for the sake of developing NLP systems that interoperate with parsers.
Proceedings of GEAF07; CSLI Publications On-line
Proceedings TOC
Proceedings as a single large pdf file