Securing IoT Malware Classifiers: Dynamic Trigger-Based Attack and Mitigation
Bibliographic record
Abstract
The evolution of IoT malware has ignited interest in the creation of malware family classification models. Nonetheless, these models encounter security concerns stemming from issues related to their interpretability and vulnerabilities exposed within the training pipeline. Recent research highlighted the limitations of learning-based malware classifiers, which are susceptible to backdoor attacks due to relying on human-engineered features to simplify the mapping from features to binary perturbations. In contrast, our study aligns with the current trajectory of the malware classification field, where we emphasize the detection of backdoor attacks targeted at models employing features extracted from within the model itself. To thoroughly assess model vulner-abilities, we have devised a dynamic trigger generation method based on sample features, which we refer to as “BENIGN”. This approach is used to contaminate and launch attacks on the model while also implementing a tailored training process to achieve specific attack objectives. Through experiments, we analyze the impact of variables involved in its training procedures on the attack stability and success rates. Last, we evaluate mitigation methods and emphasize the challenges and adaptability needed to defend against these attack strategies.
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 | 0.002 | 0.008 |
| 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 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
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".