Coca: Improving and Explaining Graph Neural Network-Based Vulnerability Detection Systems

TL;DR

Coca is a framework that enhances the robustness and explainability of GNN-based vulnerability detection systems by reducing spurious correlations and providing concise, causal explanations for detected vulnerabilities.

Contribution

The paper introduces Coca, a novel framework combining robust training and causal explanation methods to improve GNN vulnerability detectors.

Findings

Coca effectively mitigates spurious correlation issues.

Coca provides high-quality, concise explanations.

Experimental results show improved robustness and explanation quality.

Abstract

Recently, Graph Neural Network (GNN)-based vulnerability detection systems have achieved remarkable success. However, the lack of explainability poses a critical challenge to deploy black-box models in security-related domains. For this reason, several approaches have been proposed to explain the decision logic of the detection model by providing a set of crucial statements positively contributing to its predictions. Unfortunately, due to the weakly-robust detection models and suboptimal explanation strategy, they have the danger of revealing spurious correlations and redundancy issue. In this paper, we propose Coca, a general framework aiming to 1) enhance the robustness of existing GNN-based vulnerability detection models to avoid spurious explanations; and 2) provide both concise and effective explanations to reason about the detected vulnerabilities. \sysname consists of two core parts referred to as Trainer and Explainer. The former aims to train a detection model which is robust to random perturbation based on combinatorial contrastive learning, while the latter builds an explainer to derive crucial code statements that are most decisive to the detected vulnerability via dual-view causal inference as explanations. We apply Coca over three typical GNN-based vulnerability detectors. Experimental results show that Coca can effectively mitigate the spurious correlation issue, and provide more useful high-quality explanations.