DSD: Distributed Spoken Dialogue Systems

  • Brøndsted, Tom (Projektdeltager)



Speech and Multimedia Communication Focus on DSD's, e.g. remote server-based architectures where end-user's access web-services through handheld devices, have lead to standardization efforts for formalisms for recognition grammars and dialogue descriptions. Standards are discussed and proposed within the W3Cs voice browser activity, VoiceXML forum etc. Based on a flexible suite of Natural Language Processing tools developed and improved over the last years (cf. CPK Annual Report 98, 99) the project aims at evaluating some of the standard proposals in realistic a set-up, where speech grammars and dialogue descriptions are derived from changing databases to be accessed by end-users. The evaluation sets focus on core issues like ambiguity, lexical structure, and semantic representation. A demonstrator/prototype accessing news over the internet has been established. Results are reported in [Brøndsted, 2001] and [Bai and Brøndsted, 2000]. (Tom Brøndsted, Bo Nygaard Bai)
Effektiv start/slut dato31/12/200131/12/2001