Consensus-based Distributed Quantum Kernel Learning for Speech Recognition

This paper introduces the Consensus-based Distributed Quantum Kernel Learning (CDQKL) framework to advance speech recognition using quantum computing. Key contributions include:

• Distributed learning across quantum terminals connected via classical channels, ensuring data privacy by exchanging model parameters instead of raw data.
• Improved computational scalability for quantum kernel methods applied to large-scale speech emotion recognition tasks.
• Demonstrated competitive performance compared to centralized and local models, validating its potential in sectors requiring privacy-preserving ML such as telecommunications, automotive, and finance.

CDQKL represents a practical and secure approach to scaling quantum machine learning in distributed, privacy-sensitive environments.

Recommended citation: Chen, Kuan-Cheng; Ma, Wenxuan; & Xu, Xiaotian. (2025). "Consensus-based Distributed Quantum Kernel Learning for Speech Recognition." Proceedings of ICASSP 2025 – IEEE International Conference on Acoustics, Speech, and Signal Processing.
Download Paper