IGLOO SECURITY, INC. acquires a patent related to artificial intelligence, further enhances AI security monitoring efficiency

– Strengthen cyber defense by accurately selecting high-risk events 

– Continue research and development to increase AI security monitoring efficiency  

[November 19, 2020] IGLOO SECURITY, INC. (CEO Lee Deuk-choon, www.igloosec.com/en) announced that it has acquired a patent related to AI (artificial intelligence) that enhances the efficiency of transfer learning. IGLOO SECURITY is a strategy to strengthen the ability to respond to advanced cyber intrusion attempts by applying this patented technology to 'SPiDER TM AI Edition', an AI security monitoring solution. 

In order to increase the accuracy of supervised learning machine algorithms, a “labeling” process that attaches correct answers to security events is essential. As the number of security data requiring labeling has increased exponentially, “transfer learning” has been introduced, in which labels are assigned to data generated in a specific environment and re-learned according to different environments.

However, 'transfer learning' has a limitation in that the accuracy of predicted results for security events may be degraded depending on differences in security environments. This is because the judgment of a security person in charge of a specific security event may vary according to each security environment in which security data occurs, such as events deemed to be attacks in a specific security environment are considered negligible in other environments. 

The patent acquired this time by IGLOO SECURITY is focused on overcoming these difficulties and increasing the efficiency of transfer learning. This is a method of creating multiple imitation environments that imitate them through learning about environments in various conditions, and extracting and applying correction values that analyze differences between these imitation environments. Even when learning data generated in another environment is learned, a prediction result suitable for the environment in which transfer learning is performed may be derived. 

“As the digital transformation accelerates, the amount of security events that security personnel need to analyze is increasing exponentially,” said Lee Deuk-choon, CEO of IGLOO SECURITY. It is expected that by applying this patented technology, the efficiency of security monitoring can be increased to the next level by quickly and accurately selecting high-risk events that may affect companies.”