MUMBAI, India, Feb. 27 -- Intellectual Property India has published a patent application (202541132965 A) filed by Dr. Mahalingam College Of Engineering And Technnology, Pollachi, Tamil Nadu, on Dec. 29, 2025, for 'a machine learning based system and method for detecting sandbox evasion techniques in malware.'

Inventor(s) include V. Eswaramurthy; and Dr. J. Ramprasath.

The application for the patent was published on Feb. 27, under issue no. 09/2026.

According to the abstract released by the Intellectual Property India: "A Machine Learning based System and Method for Detecting Sandbox Evasion Techniques in Malware Sandbox evasion techniques have become a major problem for modern malware analysis because they let malware find and avoid virtualized or regulated environments th1,1t researchers use to study their behavior. To combat this threat, this project develops a machine learning-based Sandbox Evasion Detection System designed to identify and analyze the evasion techniques employed by malware samples during execution. The system creates a safe place for mal ware samples to run and be watched for signs of evasion, such as delays in timing, virtualization checks, environment fingerprinting, and conditional payload execution. It looks for patterns in behavior and the environment, like sequences of API calls, changes to the registry, network problems, and patterns in system resources, to find people who are trying to hide bad behavior. Random Forest, Support Vector Machine (SVM), and Gradient Boosting are examples of supervised learning algorithms that learn from labeled datasets that have both known malware samples and regular malware samples. This helps them know when someone is trying to get away from something. The system also has a logging and monitoring module that keeps track of what happens while the program is running and links those events to possible evasion signatures. Users can upload and analyze samples, see behavior reports, and get real-time detection results through a web interface built on Flask. The proposed system uses machine learning and behavioral analysis to find ways to get around sandboxes in a strong and automatic way. This makes places where you can look at malware much more reliable and helpful."

Disclaimer: Curated by HT Syndication.