Abstract
The paper focuses on cognitive biometrics and their application in user authentication. To be more precise, this work aims at investigating if the inference logic of Neural Networks can be combined with the contextual analysis of Large Language Models to create a composite metric, able to accurately reflect a user’s Keystroke Dynamic and text-based Stylometrics in response to an arbitrary visual stimulus. The created artefact showed that this strategy is viable for real-world usage, having tested the solution on real participants and industry-standard datasets. In some cases, the performance testing showed superior accuracy and speed in some performance metrics compared to other contemporary solutions that exist in this area.
Original language | English |
---|---|
Number of pages | 13 |
Journal | IEEE Transactions on Emerging Topics in Computing |
Early online date | 17 Jan 2025 |
DOIs | |
Publication status | Early online - 17 Jan 2025 |
Keywords
- Neural Networks
- Large Language Models
- Deep Learning
- Cognitive Biometrics
- Stylometrics
- Contextual Analysis