The International Computer Science Institute (ICSI) is pleased to present the following talk: Articulatory Features for Discriminative Speaker Adaptation Presented by Florian Metze of Deutsche Telekom Laboratories 11:00 a.m. on Friday, April 27th ICSI Lecture Hall 1947 Center Street, Suite 600 Berkeley, CA 94704 Abstract: This talk presents a way to perform speaker adaptation for automatic speech recognition using the stream weights in a multi-stream setup, which included acoustic models for "Articulatory Features" (AFs) such as "Rounded" or "Voiced". We present supervised speaker adaptation experiments on a spontaneous speech task and compare the above stream-based approach to conventional approaches, in which the models, and not stream combination weights, are being adapted. In the approach we present, stream weights model the importance of features such as "Voiced" for word discrimination, which offers a descriptive interpretation of the adaptation parameters. We also present results on ASR using AFs on the RT-04S "Meeting" task. For more information about ICSI, please see our website: http://www.icsi.berkeley.edu For directions to ICSI: http://www.icsi.berkeley.edu/about/location.html ----------------------------------- Leah Hitchcock ICSI Communications Coordinator leahh@icsi.berkeley.edu ICSI Seminars mailing list http://mailman.ICSI.Berkeley.EDU/mailman/listinfo/seminars-list