A Boosting Approach to Topic Spotting on Subdialogues Kary Myers (joint work with Michael Kearns, Satinder Singh, and Marilyn Walker of AT&T Labs) I'll report the results of a study we have performed on the problem of topic spotting on conversational speech. I'll describe our machine learning approach to building classifiers that accept an audio file of conversational human speech as input, and output an estimate of the topic being discussed. The methodology makes use of a well-known corpus of transcribed and topic-labeled speech (the Switchboard corpus), and involves an interesting double use of the BoosTexter learning algorithm. Our work is distinguished from previous efforts in topic spotting by an explicit study of the effects of dialogue length on classifier performance, and by our use of off-the-shelf speech recognition technology. One of our main results is the identification of a single classifier with good performance (relative to our classifier space) across all subdialogue lengths.