Unsupervised ASR Framework and Bayesian Generative Modeling Approach

Explore an unsupervised automatic speech recognition framework through a Bayesian generative modeling approach in this 10-minute conference talk presented by the Center for Language & Speech Processing at Johns Hopkins University. Delve into advanced methodologies for developing ASR systems without relying on supervised training data, examining how Bayesian statistical principles can be applied to model speech recognition processes. Learn about the theoretical foundations and practical implementations of generative models in the context of unsupervised learning for speech technology, understanding how these approaches can potentially reduce dependency on large labeled datasets while maintaining recognition accuracy.