Melola Discovery

← Melola Discovery

Akses terbuka Jurnal SINTA 1

A dilution-based defense method against poisoning attacks on deep learning systems

International Journal of Electrical and Computer Engineering (IJECE) · Vol. 14 No. 1 (2024)

DOI
10.11591/ijece.v14i1.pp645-652
ISSN
2722-2578, 2088-8708
Terbit
2024-02-01
Penerbit
Institute of Advanced Engineering and Science
Lisensi
creativecommons.org/licenses/by-sa/4.0

Abstrak

Poisoning attack in deep learning (DL) refers to a type of adversarial attack that injects maliciously manipulated data samples into a training dataset for the purpose of forcing a DL model trained based on the poisoned training dataset to misclassify inputs and thus significantly degrading its performance and reliability. Meanwhile, a traditional defense approach against poisoning attacks tries to detect poisoned data samples from the training dataset and then remove them. However, since new sophisticated attacks avoiding existing detection methods continue to emerge, a detection method alone cannot effectively counter poisoning attacks. For this reason, in this paper, we propose a novel dilution-based defense method that mitigates the effect of poisoned data by adding clean data to the training dataset. According to our experiments, our dilution-based defense technique can significantly decrease the success rate of poisoning attacks and improve classification accuracy by effectively reducing the contamination ratio of the manipulated data. Especially, our proposed method outperformed an existing defense method (Cutmix data augmentation) by 20.9%p at most in terms of classification accuracy.

Jaringan kolaborasi para penulis

3 penulis · 2 tautan kepenulisan bersama

Arahkan kursor atau ketuk sebuah nama untuk menyorot rekan penulisnya. Klik dua kali untuk membuka halaman penulis.

  • Korea National Defense University
  • Organisasi berbeda
  • Organisasi sama
  • Afiliasi tidak diketahui
  • Garis lebih tebal = lebih banyak artikel bersama · lingkaran lebih besar = lebih banyak artikel
Tampilkan sebagai tabel
Rekan penulisOrganisasiArtikel bersama
Yeongrok RahKorea National Defense University1

Afiliasi berasal dari OpenAlex; metadata OAI-PMH jurnal tidak memuatnya. Penulis tanpa ID OpenAlex dicocokkan berdasarkan nama, sehingga dua orang bisa tergabung atau satu orang bisa terpisah.

Data ini berasal dari Crossref (yang disetorkan penerbit), karena endpoint OAI-PMH milik jurnal tidak dapat di-harvest.

Sumber metadata: Crossref API (cadangan, berdasarkan ISSN) https://api.crossref.org/journals/2088-8708/works, pertama di-harvest 27 Okt 2026, terakhir terlihat 15 Nov 2026.
Metadata sebagai oai_dc, dikonversi dari Crossref
<oai_dc:dc xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:oai_dc="http://www.openarchives.org/OAI/2.0/oai_dc/"><dc:title>A dilution-based defense method against poisoning attacks on deep learning systems</dc:title><dc:creator>Park, Hweerang</dc:creator><dc:creator>Cho, Youngho</dc:creator><dc:description>&lt;jats:p&gt;Poisoning attack in deep learning (DL) refers to a type of adversarial attack that injects maliciously manipulated data samples into a training dataset for the purpose of forcing a DL model trained based on the poisoned training dataset to misclassify inputs and thus significantly degrading its performance and reliability. Meanwhile, a traditional defense approach against poisoning attacks tries to detect poisoned data samples from the training dataset and then remove them. However, since new sophisticated attacks avoiding existing detection methods continue to emerge, a detection method alone cannot effectively counter poisoning attacks. For this reason, in this paper, we propose a novel dilution-based defense method that mitigates the effect of poisoned data by adding clean data to the training dataset. According to our experiments, our dilution-based defense technique can significantly decrease the success rate of poisoning attacks and improve classification accuracy by effectively reducing the contamination ratio of the manipulated data. Especially, our proposed method outperformed an existing defense method (Cutmix data augmentation) by 20.9%p at most in terms of classification accuracy.&lt;/jats:p&gt;</dc:description><dc:publisher>Institute of Advanced Engineering and Science</dc:publisher><dc:date>2024-02-01</dc:date><dc:type>info:eu-repo/semantics/article</dc:type><dc:format>application/pdf</dc:format><dc:identifier>https://doi.org/10.11591/ijece.v14i1.pp645-652</dc:identifier><dc:identifier>10.11591/ijece.v14i1.pp645-652</dc:identifier><dc:source>International Journal of Electrical and Computer Engineering (IJECE); Vol 14 No 1 (2024); 645</dc:source><dc:source>2722-2578</dc:source><dc:source>2088-8708</dc:source><dc:relation>https://ijece.iaescore.com/index.php/IJECE/article/viewFile/32234/17122</dc:relation><dc:rights>http://creativecommons.org/licenses/by-sa/4.0</dc:rights></oai_dc:dc>