CONF bengio:2001:icassp/IDIAP Learning the Decision Function for Speaker Verification Bengio, Samy Mariéthoz, Johnny EXTERNAL https://publications.idiap.ch/attachments/reports/2000/rr00-40.pdf PUBLIC https://publications.idiap.ch/index.php/publications/showcite/bengio:2000:rr00-40 Related documents IEEE International Conference on Acoustic, Speech, and Signal Processing, ICASSP 2001 Salt Lake, City, USA IDIAP-RR 00-40 This paper explores the possibility to replace the usual thresholding decision rule of log likelihood ratios used in speaker verification systems by more complex and discriminant decision functions based for instance on Linear Regression models or Support Vector Machines. Current speaker verification systems, based on generative models such as HMMs or GMMs, can indeed easily be adapted to use such decision functions. Experiments on both text dependent and text independent tasks always yielded performance improvements and sometimes significantly. REPORT bengio:2000:rr00-40/IDIAP Learning the Decision Function for Speaker Verification Bengio, Samy Mariéthoz, Johnny EXTERNAL https://publications.idiap.ch/attachments/reports/2000/rr00-40.pdf PUBLIC Idiap-RR-40-2000 2000 IDIAP published in IEEE International Conference on Acoustic, Speech, and Signal Processing This paper explores the possibility to replace the usual thresholding decision rule of log likelihood ratios used in speaker verification systems by more complex and discriminant decision functions based for instance on Linear Regression models or Support Vector Machines. Current speaker verification systems, based on generative models such as HMMs or GMMs, can indeed easily be adapted to use such decision functions. Experiments on both text dependent and text independent tasks always yielded performance improvements and sometimes significantly.