Creating Spoken Dialogue Characters from Corpora without Annotations (bibtex)
by Gandhe, Sudeep and Traum, David
Abstract:
Virtual humans are being used in a number of applications, including simulation-based training, multi-player games, and museum kiosks. Natural language dialogue capabilities are an essential part of their human-like persona. These dialogue systems have a goal of being believable and generally have to operate within the bounds of their restricted domains. Most dialogue systems operate on a dialogue-act level and require extensive annotation efforts. Semantic annotation and rule authoring have long been known as bottlenecks for developing dialogue systems for new domains. In this paper, we investigate several dialogue models for virtual humans that are trained on an unannotated human-human corpus. These are inspired by information retrieval and work on the surface text level. We evaluate these in text-based and spoken interactions and also against the upper baseline of human-human dialogues.
Reference:
Creating Spoken Dialogue Characters from Corpora without Annotations (Gandhe, Sudeep and Traum, David), In Interspeech 2007, 2007.
Bibtex Entry:
@inproceedings{gandhe_creating_2007,
	address = {Antwerp, Belgium},
	title = {Creating {Spoken} {Dialogue} {Characters} from {Corpora} without {Annotations}},
	url = {http://ict.usc.edu/pubs/Creating%20Spoken%20Dialogue%20Characters%20from%20Corpora%20without%20Annotations%20.pdf},
	abstract = {Virtual humans are being used in a number of applications, including simulation-based training, multi-player games, and museum kiosks. Natural language dialogue capabilities are an essential part of their human-like persona. These dialogue systems have a goal of being believable and generally have to operate within the bounds of their restricted domains. Most dialogue systems operate on a dialogue-act level and require extensive annotation efforts. Semantic annotation and rule authoring have long been known as bottlenecks for developing dialogue systems for new domains. In this paper, we investigate several dialogue models for virtual humans that are trained on an unannotated human-human corpus. These are inspired by information retrieval and work on the surface text level. We evaluate these in text-based and spoken interactions and also against the upper baseline of human-human dialogues.},
	booktitle = {Interspeech 2007},
	author = {Gandhe, Sudeep and Traum, David},
	month = aug,
	year = {2007},
	keywords = {Virtual Humans}
}
Powered by bibtexbrowser