Automated message prioritization: making voicemail retrieval more efficient

ABSTRACT

Navigating through new voicemall messages to find messages of interest is a time-consuming task, particularly for high-volume users. When checking messages under a time contraint (e.g., during a brief meeting break), users need to identify those messages requiring urgent action since not all messages can be processed in limited time. For these users, it would be useful if messages of greater urgency can be played first. For other users, distinguishing personal from business voicemail is a pressing need, to separate their home and business lives. We have successfully applied machine-learning techniques to lexical, acoustic, and contextual features of voicemail in order to sort messages based on urgency and on business-relevance.

REFERENCES

• 1: Yoav Freund , Raj D. Iyer , Robert E. Schapire , Yoram Singer, An Efficient Boosting Algorithm for Combining Preferences, Proceedings of the Fifteenth International Conference on Machine Learning, p.170-178, July 24-27, 1998
• 2: Hirschberg, J., et al. SCANMail: Browsing and Searching Speech Data by Content. Proceedings of EuroSpeech, 2001.
• 3: Steve Whittaker , Julia Hirschberg , Christine H. Nakatani, All talk and all action: strategies for managing voicemail messages, CHI 98 conference summary on Human factors in computing systems, p.249-250, April 18-23, 1998, Los Angeles, California, United States [doi>10.1145/286498.286732]

INDEX TERMS

Keywords:
machine learning, phone interface, speech recognition, voicemail