Webroot Uses 5th Generation Machine Learning
For machine learning, most security vendors today use Bayesian Networks or Support Vector Machine (SVM) models to populate work queues for human analysis, which aren’t scalable or even particularly accurate
- Webroot uses Maximum Entropy Discrimination (MED) for highly accurate and scaleable web threat analysis
- Through MED, the Webroot Intelligence Network can classify 2,500+ URLs per second at an error rate of less than 2% (versus an average human rate of 5-15%)
Webroot Endpoint Shields
- Real-Time System Shield -protecting the endpoint from threats of infection
- Behavior Shield –analyzes behaviors and with the cloud, identifies malicious ones
»Core System Shield –protects the system structures from infection and malicious tampering.
- USB Shield –blocks malicious activity from removable media drives
- Infrared Shield –automatically adjusts security heuristics based upon riskiness of individual user’s behaviour
- Offline Shield –protects against persistent threats if the cloud is unavailable
- Zero-day Shield –identifies and blocks new polymorphic threats entering via exploits.
- Real-Time Anti-Phishing Shield -delivers 99%+ accuracy in identifying new phishing site -including those 'live' for seconds
