Fusing Two AI Detectors Could Catch Zero-Day Attacks on Factory Networks
A new study shows that adaptively fusing a closed-world classifier with an unsupervised open-set detector significantly improves detection of zero-day ...
A new study shows that adaptively fusing a closed-world classifier with an unsupervised open-set detector significantly improves detection of zero-day ...
A new systematic review in Artificial Intelligence Review maps how multi-objective optimization algorithms are being used to balance accuracy, false ...
Researchers in Vietnam show that multimodal intrusion detection systems hardened with adversarial training can resist transferable adversarial examples generated by ...
Researchers in Vietnam have developed a Dual SPLIR epidemic-style model that incorporates APT dormancy, network heterogeneity, and organizational policy to ...
A new hybrid framework combines elliptic-curve cryptography with quantum-resistant algorithms to defend cloud infrastructure against DDoS attacks.
Researchers have developed an immune cooperative hybrid breeding optimization algorithm that selects features more effectively for intrusion detection, outperforming 13 ...
Czech researchers have unveiled a Siamese neural network approach that identifies encrypted video streams with near state-of-the-art accuracy while eliminating ...
Researchers have developed a hybrid Greater Cane Mayfly Algorithm that improves routing, quality of service, and resource efficiency in mobile ...
An open-source simulation framework called DDoSimu5G generates precisely labeled DDoS attack datasets inside a simulated 5G network, giving machine-learning intrusion ...
Researchers have developed an autonomous blockchain-based system that improves intrusion detection accuracy while cutting latency and energy consumption in fog ...
© 2025 Scienmag - Science Magazine
© 2025 Scienmag - Science Magazine