Securayer Registers Patent for 'Anomaly Web Log Detection Technology'

Securayer has registered a patent for an unsupervised learning method and device that detects anomaly web logs using imbalanced training data. (From left: Patent inventors Lee Geon-hyeok, Kim Sang-hyun, and Kim Jin, researchers at Securayer)

Securayer is enhancing its security competitiveness by developing technology that can detect cyber attacks even in situations of data imbalance.

On the 9th, Securayer announced that it has registered a patent containing methods and learning devices for detecting anomaly web logs from imbalanced data using artificial intelligence technology.

This patent focuses on implementing a deep learning model based on unsupervised learning to solve the problem of imbalanced data. It also provides a method for clustering feature values using clustering operations and determining data with those features as anomalous data.

Kim Sang-hyun, a researcher at Securayer, stated, “As remote work environments increase and digital transformation accelerates, various forms of cyber attacks are on the rise,” and added, “The commonly used ‘rule-based anomaly detection method’ is vulnerable to new types of attacks.”

He further explained, “While ‘supervised learning-based detection methods’ are being developed as an alternative, there are limitations due to difficulties in learning caused by data shortages when using new data,” emphasizing, “We have developed technology to detect anomaly web logs even in situations of data imbalance by utilizing unsupervised learning-based deep learning models and clustering operations, which will facilitate the detection of new types of cyber attacks in the future.”

Meanwhile, this patent was applied for and registered as part of the ‘SW Computing Industry Source Technology Development Project (2020-0-00107)’ supported by the Institute for Information & Communications Technology Planning & Evaluation (IITP).

Source: Edaily (www.edaily.co.kr)