An Extension to VoiceXML With Semantic Tagging Support and its Application

谢伟凯,史元春,徐光祐
DOI: https://doi.org/10.3969/j.issn.1000-1220.2002.12.002
2002-01-01
Abstract:We present in this paper a practical semantic tagging scheme based on VoiceXML, and an automatic method to generate the corresponding semantic interpreters is described as well. The extended VoiceXML can not only specify syntax of the voice interface in a speech-enabled system, but also specify the associated semantic actions. Our semantic tagging scheme demonstrates more expressive capability and flexibility, compared to other schemes provided by speech recognition engine vendors used in their proprietary APIs. By adopting this kind of extended VoiceXML in a speech-enabled system, the voice interface specification and the user agent logic that execute tasks on behalf of the user can be clearly separated, and therefore make the system more flexible and improve the develop efficiency.
What problem does this paper attempt to address?