Google, IBM & Meta Certificates — All 10,000+ Courses at 40% Off
One annual plan covers every course and certificate on Coursera. 40% off for a limited time.
Get Full Access
Explore sequence kernels for speaker and speech recognition in this 2009 lecture by Mark Gales from the University of Cambridge, presented at the Center for Language & Speech Processing at Johns Hopkins University. Delve into the concept of sequence kernels and their application in mapping variable-length sequences to fixed-dimensional feature spaces, making them suitable for time-varying speech signals. Examine the successful use of sequence kernels in speaker verification, particularly when combined with support vector machines (SVMs) for classification. Focus on generative kernels, a specific class of sequence kernels, and their utilization in both speaker and speech recognition. Investigate how generative kernels and score-spaces leverage generative models like hidden Markov models (HMMs) and Gaussian mixture models (GMMs) to extract fixed-dimensional feature vectors. Learn about the Fisher Kernel and its applications in biological sequences, as well as its relationship to the GMM mean-Supervector kernel commonly used in speaker verification. Discover how these kernels and associated feature spaces can be applied to speech recognition and address speaker and environment changes.
