System2Vec
Large-Scale Security Management Based on Deep Learning
Autonomic Cyber Threat Classification Based on Deep Learning
Models 800 million threats from 1,000 TMS devices based on Deep Learning
Detects intrusion attepts, Web hacking, tunneling exploitation, malware, DDoS, and hacking Emails
Highly accurate classification with an f1-score of 99.8%
Classification Result Interpretation with XAI Technologies
Pinpoints attack factors using Transformer and attention mechanism that are used by ChatGPT as well
Highlights the parts of network payload suspected to be malicious
Precise quantification of maliciousness
Recommendation of Detection Rule Revisions
Recommends Snort rules based on our XAI's interpretation of cyber threat classification results
Reduced false positives based on automatically improved detection rules
Context-awareness
Correlation between different cyber threats
Association rule mining of network flows and cyber threats to identify abnormal symptoms and prognosis
Intuitive detection of abnormal network behaviors with complex network features
Highly Portable, Efficient, and
Cloud-Ready!
Cloud-Ready!
Supports CI/CD (Continuous Integration / Continuous Deployment) for timely update of AI models
Capable of processing 150,000 cyber threats per minute with commodity GPUs
Runs on individual TMS/IDS devices or on Cloud environments
Sustainable Machine Learning Quality
High accuracy proved with CICIDS2017 open data set and threat data collected at the Education Cyber Security Center in Korea
Deals with seasonal changes in threat patterns through continuous fine-tuning of the AI model
System2Vec, who is it for?
Company or public organizations exposed to many constant threats
Those who cannot afford human agents for security monitoring
Those who suffer from inaccurate TMS/IDS devices