Hearing Without Listening: Voice Authentication with Privacy

Explore the critical issue of privacy in speech processing systems through this insightful lecture by Bhiksha Raj from Carnegie Mellon University. Delve into the challenges of voice authentication systems, where voice prints can be vulnerable to theft or unauthorized use. Learn about innovative approaches to address these privacy concerns, including cryptographic methods and secure binary embeddings. Discover how these techniques aim to enable voice authentication without storing interpretable voice prints, potentially revolutionizing the field of speech processing. Gain valuable insights into the intersection of speech technology, privacy, and security from an expert in automatic speech recognition, audio processing, and machine learning.