Machine learning system that detects software vulnerabilities by converting source code to vector representations, generating images, and classifying them using trained neural networks.

Patentradars sammanfattning

The invention relates to automated vulnerability detection in source code using machine learning. The system transforms source code into vector representations, generates corresponding images, resizes them to a uniform dimension, and applies a pre-trained ML model to classify the code as either vulnerability-free or containing one or more vulnerabilities. This image-based classification approach leverages visual pattern recognition to identify security flaws in software.

AI-genererad och redaktionellt processad.