Zenedge introduces machine learning cyber security solution
Zenedge, cloud-based cyber security solutions provider, has introduced a Web application firewall that leverages artificial intelligence to update security postures to protect Web applications from vulnerabilities.
According to the company, leveraging Zenedge's patent-pending mathematical model based on proprietary machine learning algorithms and big data analysis, Zenedge AI inspects Web traffic in real-time, identifies threats and behaviour anomalies, and updates security postures.
Leon Kuperman, CTO of Zenedge, says there's no universal mathematical model to describe malicious user behaviour, which is why the signature-based model used by many legacy Web application security vendors is so prone to false positives.
"By combining machine learning algorithms with a data-driven model in Zenedge AI, our platform can now detect cyber threats and behaviour anomalies in real-time, and update security postures dynamically without any need for human intervention."
He says on the Web application front, organisations across the board are subjected to cyber attacks more and more often, and enterprises can no longer afford to remain vulnerable until a patch exists.
Enterprises need to break free from their reliance on the standard signature-based approach to Web application security as an effective defence against zero-day attacks, notes Kuperman, adding this method requires constant tuning and significant human resources.
"As threat actors and cyber attacks become increasingly more sophisticated and new 'zero day' attempts are made to find vulnerabilities in Web sites and Web applications, enterprises are often left with a large window of exposure as these emerging threats go undetected or unresolved until a patch is available."
By leveraging artificial intelligence to do this, the rampant industry problem of false positives is negated, says Kuperman.
The Zenedge AI platform is continuously "learning" what malicious and legitimate behaviours look like and identifying patterns in real-time with respect to each enterprise's Web site and applications, as no single model is applicable across all, he adds.
Common Web application attacks such as XSS and SQL injection can be quickly separated out, as every new request that comes in is then evaluated against the model, says Kuperman.
This better equips enterprises with the ability to identify and mitigate zero-day threats quickly and effectively, he concludes.