Calibrating neural networks for synthetic speech detection: A likelihood-ratio-based approach

Cuccovillo L, Aichroth P, Köllmer T. Calibrating neural networks for synthetic speech detection: A likelihood-ratio-based approach. In: AES 2024 International Conference on Audio Forensics. Audio Engineering Society; 2024. Paper 3. Available from: https://aes.org/publications/elibrary-page/?id=22628

@inproceedings{Cuccovillo2024_22628,
  author = {Cuccovillo, Luca and Aichroth, Patrick and Köllmer, Thomas},
  title = {{Calibrating neural networks for synthetic speech detection: A likelihood-ratio-based approach}},
  booktitle = {AES 2024 International Conference on Audio Forensics},
  note = {Paper 3},
  year = {2024},
  month = jun,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=22628}
}

TY  - CPAPER
TI  - Calibrating neural networks for synthetic speech detection: A likelihood-ratio-based approach
AU  - Cuccovillo, Luca
AU  - Aichroth, Patrick
AU  - Köllmer, Thomas
T2  - AES 2024 International Conference on Audio Forensics
M1  - Paper 3
PY  - 2024
DA  - 2024/06/17
UR  - https://aes.org/publications/elibrary-page/?id=22628
PB  - Audio Engineering Society
LA  - en
AB  - In this paper, we introduce a calibration procedure designed to convert the uncalibrated output scores of neural networks for synthetic speech detection into calibrated and interpretable likelihood ratios. This procedure is based on the assumption that the networks subject to calibration are deterministic and have undergone training until they reached convergence. Provided these conditions are satisfied, it is then possible to transform their output values into likelihood ratios using a minimal set of validation and calibration data, eliminating the need for retraining the models. We successfully tested the entire workflow on a state-of-the-art network example, demonstrating not only its effectiveness in calibration but also its ability to enhance fault tolerance against inadequate inputs.
ER  -