PIPE gDAO

Threat Detection and Response with Heterogeneous Data Sources

Threat Hunting
Hidden Markov Model
Time Series Decomposition
Beaconing
Botnet
Adversarial Evasion
Sporadic communication

Reveal complex patterns in beacon messages despite the absence of labelled data, detecting and mitigating botnets and their beaconing activities.

project image

Project Details

Project Timeline

Preparing

Disclosing

Being Reviewed

Reporting

Decision

The statement discusses the development of an unsupervised multi-model called NetSpectra Sentinel (NSS) for identifying threats hidden within benign applications' activities, utilizing Continuous-Time Hidden Markov Model (CT-HMM) and Time Series Decomposition (TSD) to uncover hidden patterns in system network logs. NSS can effectively reveal complex patterns in beacon messages despite the absence of labeled data. It highlights the challenges in detecting and mitigating botnets and their beaconing activities, explaining the concept of botnets, beaconing, and the significance of Distributed Denial of Service (DDoS) attacks. The focus is on addressing the challenge of Attacker-Driven Beaconing.