A Novel Data-Driven Attack Method on Machine Learning Models
Yazarlar (3)
Arş. Gör. Emre SADIKOĞLU Yalova Üniversitesi, Türkiye
Dr. Öğr. Üyesi İrfan Kösesoy Kocaeli Üniversitesi, Türkiye
Prof. Dr. Murat GÖK Yalova Üniversitesi, Türkiye
| Makale Türü |
Özgün Makale
(SSCI, AHCI, SCI, SCI-Exp dergilerinde yayınlanan tam makale)
|
||
| Dergi Adı | Journal of Universal Computer Science (Q3) | ||
| Dergi ISSN | 0948-695X Dergi Bilgileri (2024) | ||
| Dergi Tarandığı Indeksler | SCI-Expanded | ||
| Makale Dili | Türkçe | Basım Tarihi | 03-2024 |
| Kabul Tarihi | – | Yayınlanma Tarihi | 28-03-2024 |
| Cilt / Sayı / Sayfa | 30 / 3 / 402–417 | DOI | 10.3897/jucs.108445 |
| Makale Linki | https://doi.org/10.3897/jucs.108445 | ||
| UAK Araştırma Alanları |
Siber Güvenlik
Makine Öğrenmesi
Yapay Zeka
|
||
| Özet |
| With the increasing popularity and usage of artificial intelligence systems, it has become crucial to address their vulnerability to cyber-attacks. In this study, we propose a novel gradient descent-based method to generate fake data that can be accepted as positive by a targeted machine learning model. Our method is designed to generate a large number of positive samples with a minimal number of probes to the model, making it difficult to detect by security systems. Additionally, we develop an alternative model to the attacked model using a reverse engineering approach, trained on a dataset composed of the samples generated by our method. We evaluate the success of our proposed method and the alternative model through a series of experiments. We conducted experiments on six distinct datasets, each of which was trained using three separate machine-learning algorithms. This resulted in a total of eighteen … |
| Anahtar Kelimeler |
| Adversarial data | Artificial intelligence | Cyber security | Data driven attack |
BM Sürdürülebilir Kalkınma Amaçları
| Atıf Sayıları | |
| Yöksis | 1 |
| Web of Science | 1 |
| Scopus | 2 |
| Google Scholar | 2 |