Transfer Learning of Keystroke Dynamics for Cross-Device User Authentication

2026-08-17

AI Summary

This study addresses the distribution drift caused by device form factor variations in cross-device keystroke authentication. We propose an inductive transfer learning-based framework that leverages feature adaptation and an extended feature set to enhance model discriminability under limited target device data, effectively mitigating challenges arising from typing pattern shifts. Experimental evaluations on the BBMAs dataset demonstrate that the proposed system achieves a cross-device equal error rate of 14.2%, significantly outperforming state-of-the-art methods. These results confirm the framework’s efficacy in overcoming domain discrepancies, offering a robust and efficient solution for cross-device biometric authentication in heterogeneous environments.

Abstract

Keystroke dynamics (typing patterns) can be used as a behavioural biometric modality for user authentication, with applications such as fraud prevention. While the modality has been shown to work well for single device authentication, its application to cross-device scenarios is more challenging. Dynamics learned on one device (eg., phone) may not be directly applicable to authentication on a secondary device with a different form factor (eg., tablet) due to changes in typing patterns that can lead to distribution drifts. To address this, we propose a cross-device user authentication system based on inductive transfer learning, where keystroke dynamics learned on one device are adapted to a secondary device. The adapted data is then combined with necessarily limited training data for the secondary device, which is used to robustly train a binary classifier. Furthermore, an extended set of keystroke features is used to better capture discriminative dynamics. Experiments on the BBMAS dataset show that proposed system achieves an equal error rate of 14.2% for the cross-device scenario, surpassing state-of-the-art methods.

Problem

• Keystroke Dynamics
• Cross-Device Authentication
• Distribution Drift
• Behavioral Biometrics

Innovation

• Inductive Transfer Learning
• Cross-Device Authentication
• Keystroke Dynamics
• Distribution Drift
• Extended Feature Set