Speech and Dialog Mining on Heterogeneous Data

Explore speech and dialog mining techniques for heterogeneous data in this informative lecture by Allen Gorin from AT&T Research. Delve into the critical aspects of customer interactions in business, focusing on both human agent and automated system communications. Examine current methods for analyzing, searching, and acting upon customer interactions, highlighting their labor-intensive nature and limitations. Discover innovative research aimed at enabling businesses to efficiently browse, prioritize, select, and extract information from vast volumes of customer interactions. Address the key technical challenge of working with heterogeneous data, which combines speech and associated call/caller information. Gain insights from experimental evaluations conducted on AT&T's 'How May I Help You?'(sm) spoken dialog system, demonstrating practical applications of these advanced mining techniques.