Executable specifications of natural-language speech applications

点击翻译按钮获取中文摘要

The functionality of computers and the world wide web would be significantly improved if there were more theories and open-source tools that could be used by non-experts to build and deploy natural-language speech applications. Not only would this benefit people with visual, cognitive, or motor disabilities, it would also benefit users who require hands-free access. The objective of this research is 1) to investigate various semantic theories of natural-language and to develop a compositional semantics that will facilitate the construction of efficient user-friendly natural-language interfaces to databases and other sources of knowledge on the web, 2) to develop theory and tools which will allow natural-language processors to be created by programmers with little knowledge of parsing technology, as executable specifications of attribute grammars which define the syntax and the semantics of the language to be processed, and 3) to create tools that will facilitate the deployment of hyperlinked natural-language speech applications that can be accessed by the public using freely-available speech browsers and commonly-used communication protocols.