What question did this study set out to answer?

The research aims to develop a lightweight model for identifying specific memory safety vulnerabilities in code.

March 4, 2026Open Access

Lightweight Memory Safety Vulnerability Detection via CodeBERT and Focal Loss: A Hybrid Training Approach on Consumer-Grade Hardware

NCNolan ChenNational Institutes of Health

Key Points

The research aims to develop a lightweight model for identifying specific memory safety vulnerabilities in code.
Developed a detection model based on CodeBERT and Focal Loss.
Implemented a hybrid training strategy using Juliet Test Suite and PrimeVul.
Conducted experiments on the CASTLE-C250 dataset with a consumer-grade GPU.
Achieved 76.67% Recall on the detection model.
Demonstrated a 4.4x improvement over the existing CodeBERT baseline.
Successfully ran the model on a consumer-grade Intel Arc A380 GPU.

Abstract

This study proposes a lightweight automated detection model for CWE-119 (Buffer Overflow) and CWE-787 (Out-of-Bounds Write) memory safety vulnerabilities, built upon CodeBERT and Focal Loss. Training adopts a hybrid strategy combining Juliet Test Suite and PrimeVul. On CASTLE-C250, the model achieves 76.67% Recall — a 4.4x improvement over the CodeBERT baseline — without graph-based structural augmentation. The complete pipeline runs on a consumer-grade Intel Arc A380 GPU using BF16 mixed precision.

AI से पूछें

Bookmark

View Full Paper