Call Classification with Hundreds of Classes and Hundred Thousands of Training Utterances ... ... and No Target Domain Data

Authors:
David Suendermann;Phillip Hunter;Roberto Pieraccini
Affiliations:
SpeechCycle, Inc., New York City, USA;SpeechCycle, Inc., New York City, USA;SpeechCycle, Inc., New York City, USA
Venue:
PIT '08 Proceedings of the 4th IEEE tutorial and research workshop on Perception and Interactive Technologies for Speech-Based Systems: Perception in Multimodal Dialogue Systems
Year:
2008

Citing 2
Cited 0

How may I help you?

Speech Communication - Special issue on interactive voice technology for telecommunication applications (IVITA '96)
Technical support dialog systems: issues, problems, and solutions

NAACL-HLT-Dialog '07 Proceedings of the Workshop on Bridging the Gap: Academic and Industrial Research in Dialog Technologies

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper reports about an effort to build a large-scale call router able to reliably distinguish among 250 call reasons. Because training data from the specific application (Target) domain was not available, the statistical classifier was built using more than 300,000 transcribed and annotated utterances from related, but different, domains. Several tuning cycles including three re-annotation rounds, in-lab data recording, bag-of-words-based consistency cleaning, and recognition parameter optimization improved the classifier accuracy from 32% to a performance clearly above 70%.