Robustness of deep autoencoder in intrusion detection under adversarial contamination
Bibliographic record
Abstract
The existing state-of-the-art in the field of intrusion detection systems (IDSs) generally involves some use of machine learning algorithms. However, the computer security community is growing increasingly aware that a sophisticated adversary could target the learning module of these IDSs in order to circumvent future detections. Consequently, going forward, robustness of machine-learning based IDSs against adversarial manipulation (i.e., poisoning) will be the key factor for the overall success of these systems in the real world. In our work, we focus on adaptive IDSs that use anomaly-based detection to identify malicious activities in an information system. To be able to evaluate the susceptibility of these IDSs to deliberate adversarial poisoning, we have developed a novel framework for their performance testing under adversarial contamination. We have also studied the viability of using deep autoencoders in the detection of anomalies in adaptive IDSs, as well as their overall robustness against adversarial poisoning. Our experimental results show that our proposed autoencoder-based IDS outperforms a generic PCA-based counterpart by more than 15% in terms of detection accuracy. The obtained results concerning the detection ability of the deep autoencoder IDS under adversarial contamination, compared to that of the PCA-based IDS, are also encouraging, with the deep autoencoder IDS maintaining a more stable detection in parallel to limiting the contamination of its training dataset to just bellow 2%.
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 | 0.003 | 0.014 |
| 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".