Learning from Execution Graphs: A Graph Neural Network Approach to Malware Detection
Bibliographic record
Abstract
The growing sophistication and scale of malicious software demand detection methods capable of capturing both the behavioral complexity and structural relationships in program execution. Graph Neural Networks (GNNs) have shown strong potential for malware detection by leveraging graph-structured representations such as Control Flow Graphs (CFGs), which preserve execution semantics and interdependencies between program components. When extracted dynamically, CFGs reflect actual runtime behavior, offering richer and more reliable information than static analysis alone. At the same time, GNN performance is sensitive to hyperparameter configurations, where suboptimal settings can lead to over-smoothing, reduced accuracy, or excessive computational cost. This paper presents a systematic study that combines the development of a GNN-based malware detection framework using dynamically extracted CFGs with a detailed analysis of hyperparameter sensitivity. We investigate the influence of model depth, hidden dimensions, learning rate, weight decay, and training epochs using two representative architectures, Graph Convolutional Networks (GCN) and GraphSAGE. The findings provide empirical insights and practical guidelines for building powerful, efficient, and well-tuned GNN-based malware detection systems, with applicability to other graph classification domains.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.007 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".