ADVERT: An Adaptive and Data-Driven Attention Enhancement Mechanism for Phishing Prevention

Attacks exploiting the innate and acquired vulnerabilities of human users have posed severe threats to cybersecurity. This work proposes ADVERT, a human-technical solution that generates adaptive visual aids in real-time prevent from inadvertence reduce their susceptibility phishing attacks. Based on eye-tracking data, we extract visual states attention as system-level sufficient statistics characterize user’s behaviors attention status. By adopting data-driven approach two learning feedback different time scales, this lays out theoretical foundation analyze, evaluate, particularly modify humans’ processes while they vet recognize emails. We corroborate effectiveness, efficiency, robustness ADVERT through case study based data set collected subject experiments conducted at New York University. The results show can statistically increase level improve accuracy recognition 74.6% minimum 86%. meta-adaptation further 91.5% (resp. 93.7%) less than 3 50) tuning stages.