AI & ML

Malware Detection and Classification

This project examines malware detection and classification directly from raw executable byte sequences, avoiding heavy manual feature engineering.

Back to Research
Malware Detection and Classification project illustration.

Research Area

AI & ML

Key Methods

Transformers, Cybersecurity, Byte-Level Modeling

Project Summary

Executable byte-sequence analysis to detect malicious patterns and improve cybersecurity response.

Traditional malware analysis often depends on handcrafted features or costly preprocessing, which can miss important byte-level structure in executable files. Identify malicious patterns directly from raw byte streams and improve malware-family classification robustness. The project analyzes raw byte sequences of executable files such as Windows PE binaries to detect suspicious behavior. The MalFormer direction treats malware analysis as a sequence modeling problem over binary content. The overall aim is to strengthen detection quality while reducing dependence on manual feature design.