EERLoss: A Novel Loss Function for Training Deep Biometric Models

This paper introduces EERLoss, a subdifferentiable approximation of the Equal Error Rate (EER) designed to align deep biometric model training with primary evaluation metrics. Validated on keystroke dynamics verification using the KVC-onGoing benchmark, the approach addresses the misalignment between optimization objectives and performance assessment.

• Evaluated on the large-scale KVC-onGoing benchmark with data from over 185,000 subjects.
• Demonstrates superiority over existing state-of-the-art loss functions through comprehensive ablation studies.
• Converges substantially faster than other losses, reducing overall training costs.
• Achieves a relative EER reduction of approximately 30% when applied to the KVC-winning architecture.

The results validate EERLoss as an effective, task-aligned training objective specifically suited for high-variance biometric traits like keystroke dynamics.