Generative Modelling for Unsupervised Score Calibration
This work addresses the need for cost-effective calibration in speaker recognition systems by eliminating the requirement for supervised data, though it appears incremental as it builds on existing calibration methods.
The paper tackled the problem of unsupervised score calibration for speaker recognition, which traditionally requires expensive supervised data, by proposing a 2-component GMM and demonstrated good performance relative to a supervised baseline on NIST SRE'10 and SRE'12 datasets.
Score calibration enables automatic speaker recognizers to make cost-effective accept / reject decisions. Traditional calibration requires supervised data, which is an expensive resource. We propose a 2-component GMM for unsupervised calibration and demonstrate good performance relative to a supervised baseline on NIST SRE'10 and SRE'12. A Bayesian analysis demonstrates that the uncertainty associated with the unsupervised calibration parameter estimates is surprisingly small.