Kaspersky Increases APT Detection by a Quarter
The cybersecurity service provider Kaspersky reports a 25 percent increase in advanced persistent threat (APT) detection during the first half of 2024, attributing this improvement to their enhanced machine learning capabilities.
Thousands of new threats were identified, with targeted attacks particularly affecting government authorities, financial institutions, and telecommunications providers. According to Kaspersky, the implementation of machine learning not only improves technical detection capabilities but also transforms subsequent threat management: the combination of automation and intelligent data analysis enables more efficient resource allocation and more effective protective measures.
To achieve these results, Kaspersky employs various analytical approaches. These include the Random Forest algorithm for classification and Term Frequency-Inverse Document Frequency (TF-IDF) for text analysis and feature extraction. These techniques process vast amounts of data to enable faster and more accurate threat detection. According to the company, this combined approach helps identify Indicators of Compromise (IoCs) that might be missed by traditional detection systems.
Data-Driven Model Updates
Kaspersky reports processing millions of data points daily, enabling near real-time threat detection and response capabilities. When new attack vectors emerge, the machine learning systems are updated under human oversight to provide fresh insights, enhance defensive capabilities, and strengthen organizational resilience.
'The results have exceeded our expectations,' says Amin Hasbini, Head of Cybersecurity Research and Global Research and Analysis Team (GReAT) at Kaspersky Middle East, Turkey and Africa. 'These technologies improve detection accuracy and promote proactive defence strategies that help businesses stay ahead of evolving cyber threats. The future of cybersecurity lies in the ethical use of these tools to build a safer world for all.'
