A gradient-based approach for adversarial attack on deep learning-based network intrusion detection systems
Bibliographic record
Abstract
Intrusion detection systems are an essential part of any cybersecurity architecture. These systems are critical in defending networks against a variety of security threats. In recent years, deep neural networks have proved their performance and efficiency in various machine learning tasks, including intrusion detection . However, it is shown that deep learning models are highly vulnerable to adversarial attacks . This paper proposes a new approach for performing an adversarial attack against deep learning-based malicious network activity classification . We use the Jacobian Saliency Map to find the best group of features, with different features and perturbation magnitude, to generate adversarial examples . We evaluate our method on three CIC-IDS2017, CIC-IDS2018, and CIC-DDoS2019 datasets. Our experiments show that our proposed method can achieve better performance while using fewer features in adversarial sample generation than other attacks that depend on a higher number of features. Our technique can generate adversarial samples for more than 18% of samples in CIC-IDS2017, 15% of samples in CIC-IDS2018, and 14% of samples in CIC-DDoS2019, using only three features and 0.1 as the perturbation magnitude. We do a deeper analysis of the attack based on its parameters, distance metrics, and the target model performance. Also, an evaluation model with three criteria, including success rates of the best feature sets, average confidence of the adversarial class, and adversarial samples transferability, is used in our analysis.
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 | 0.003 |
| 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 | 0.002 |
| 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".