Kolloquiumsvortrag 25. November 2025, Babar Ali (Betreuer: Al Sardy/Muhammad)
Dynamic Anomaly Detection for Evolving Cyber Threats Using Adaptive LSTM Networks
Cybersecurity systems must adapt to constantly evolving threats, where traditional static models, such as LSTM, struggle with concept drift and unseen attack vectors. This thesis proposes an adaptive anomaly detection system using Adaptive LSTM networks capable of adaptive learning from streaming data batches. By integrating memory replay and incremental updates, the model addresses catastrophic forgetting while improving resilience against drifts and adversarial attacks. The goal is to enhance detection accuracy and robustness compared to conventional static models in dynamic cyber environments.
Uhrzeit: 11:00 Uhr
Ort: Raum 04.137, Martensstr. 3, Erlangen
oder
Zoom-Meeting beitreten:
https://fau.zoom-x.de/j/68350702053?pwd=UkF3aXY0QUdjeSsyR0tyRWtLQ0hYUT09
Meeting-ID: 683 5070 2053
Kenncode: 647333