A GNN-Based Adversarial Internet of Things Malware Detection Framework for Critical Infrastructure: Studying Gafgyt, Mirai, and Tsunami Campaigns
Bibliographic record
Abstract
Significant advancement in Deep learning (DL) has turned it into an integral part of robust approaches for addressing cybersecurity problems in both current and aging infrastructures. Control Flow Graphs (CFGs) have demonstrated their effectiveness as leading choices that result in high-performing classifiers among various data representations used by DL-based models. Recently, Graph Neural Networks (GNNs) have made breakthroughs in the graph domain, and before long, they were jointly used with CFGs to train performant malware classifiers. However, graph-based adversarial attacks have caused suspicion about the predictions these graph-based malware classifiers make, and few studies have investigated detecting such attacks. Therefore, this paper proposes a novel GNN-based adversarial detector for identifying adversarial CFGs with higher efficacy than the previous work. This adversarial detector is placed in a data pipeline before a GNN-based malware classifier. In this paper, we solve the adversarial detection problem as an anomaly detection scenario and train the adversarial detector to learn the normal data distribution. Our GNN-based adversarial detector detects 98.96% of all adversarial CFGs, which is 1.17% higher than the previous method, with a 5.95% lower False Positive Rate (FPR). In the most hazardous category of the attack, where the attacker intends to render a malicious example as a benign input, we achieve a 4.85% boost compared to the previous competitors.
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 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 source (direct Gemma or distilled Codex), 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".