A decade of adversarial examples: a survey on the nature and understanding of neural network non-robustness
Adversarial examples, in the context of computer vision, are inputs deliberately crafted to deceive or mislead artificial neural networks. These examples exploit vulnerabilities in neural networks, resulting in minimal alterations to the original input that are imperceptible by humans but can signif...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Samara National Research University
2025-04-01
|
| Series: | Компьютерная оптика |
| Subjects: | |
| Online Access: | https://computeroptics.ru/KO/Annot/KO49-2/490209.html |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849232865090338816 |
|---|---|
| author | A.V. Trusov E.E. Limonova V.V. Arlazarov |
| author_facet | A.V. Trusov E.E. Limonova V.V. Arlazarov |
| author_sort | A.V. Trusov |
| collection | DOAJ |
| description | Adversarial examples, in the context of computer vision, are inputs deliberately crafted to deceive or mislead artificial neural networks. These examples exploit vulnerabilities in neural networks, resulting in minimal alterations to the original input that are imperceptible by humans but can significantly impact the network’s output. In this paper, we present a thorough survey of research on adversarial examples, with a primary focus on their impact on neural network classifiers. We closely examine the theoretical capabilities and limitations of artificial neural networks. After that, we explore the discovery and evolution of adversarial examples, starting from basic gradient-based techniques and progressing toward the recent trend of employing generative neural networks for this purpose. We discuss the limited effectiveness of existing countermeasures against adversarial examples. Furthermore, we emphasize that the adversarial examples originate the misalignment between human and neural network decision-making processes. That can be attributed to the current methodology for training neural networks. We also argue that the commonly used term “attack on neural networks” is misleading when discussing adversarial deep learning. Through this paper, our objective is to provide a comprehensive overview of adversarial examples and inspire further researchers to develop more robust neural networks. Such networks will align better with human decision-making processes and enhance the security and reliability of computer vision systems in practical applications. |
| format | Article |
| id | doaj-art-de74100092c84c49b0934ff28bc7a8e5 |
| institution | Kabale University |
| issn | 0134-2452 2412-6179 |
| language | English |
| publishDate | 2025-04-01 |
| publisher | Samara National Research University |
| record_format | Article |
| series | Компьютерная оптика |
| spelling | doaj-art-de74100092c84c49b0934ff28bc7a8e52025-08-20T17:16:37ZengSamara National Research UniversityКомпьютерная оптика0134-24522412-61792025-04-0149222225210.18287/2412-6179-CO-1494A decade of adversarial examples: a survey on the nature and understanding of neural network non-robustnessA.V. Trusov0E.E. Limonova1V.V. Arlazarov2Federal Research Center “Computer Science and Control” of the Russian Academy of Sciences; Smart Engines Service LLC; Moscow Institute of Physics and TechnologyFederal Research Center “Computer Science and Control” of the Russian Academy of Sciences; Smart Engines Service LLCFederal Research Center “Computer Science and Control” of the Russian Academy of Sciences; Smart Engines Service LLCAdversarial examples, in the context of computer vision, are inputs deliberately crafted to deceive or mislead artificial neural networks. These examples exploit vulnerabilities in neural networks, resulting in minimal alterations to the original input that are imperceptible by humans but can significantly impact the network’s output. In this paper, we present a thorough survey of research on adversarial examples, with a primary focus on their impact on neural network classifiers. We closely examine the theoretical capabilities and limitations of artificial neural networks. After that, we explore the discovery and evolution of adversarial examples, starting from basic gradient-based techniques and progressing toward the recent trend of employing generative neural networks for this purpose. We discuss the limited effectiveness of existing countermeasures against adversarial examples. Furthermore, we emphasize that the adversarial examples originate the misalignment between human and neural network decision-making processes. That can be attributed to the current methodology for training neural networks. We also argue that the commonly used term “attack on neural networks” is misleading when discussing adversarial deep learning. Through this paper, our objective is to provide a comprehensive overview of adversarial examples and inspire further researchers to develop more robust neural networks. Such networks will align better with human decision-making processes and enhance the security and reliability of computer vision systems in practical applications.https://computeroptics.ru/KO/Annot/KO49-2/490209.htmladversarial examplesadversarial deep learningneural networksneural network security |
| spellingShingle | A.V. Trusov E.E. Limonova V.V. Arlazarov A decade of adversarial examples: a survey on the nature and understanding of neural network non-robustness Компьютерная оптика adversarial examples adversarial deep learning neural networks neural network security |
| title | A decade of adversarial examples: a survey on the nature and understanding of neural network non-robustness |
| title_full | A decade of adversarial examples: a survey on the nature and understanding of neural network non-robustness |
| title_fullStr | A decade of adversarial examples: a survey on the nature and understanding of neural network non-robustness |
| title_full_unstemmed | A decade of adversarial examples: a survey on the nature and understanding of neural network non-robustness |
| title_short | A decade of adversarial examples: a survey on the nature and understanding of neural network non-robustness |
| title_sort | decade of adversarial examples a survey on the nature and understanding of neural network non robustness |
| topic | adversarial examples adversarial deep learning neural networks neural network security |
| url | https://computeroptics.ru/KO/Annot/KO49-2/490209.html |
| work_keys_str_mv | AT avtrusov adecadeofadversarialexamplesasurveyonthenatureandunderstandingofneuralnetworknonrobustness AT eelimonova adecadeofadversarialexamplesasurveyonthenatureandunderstandingofneuralnetworknonrobustness AT vvarlazarov adecadeofadversarialexamplesasurveyonthenatureandunderstandingofneuralnetworknonrobustness AT avtrusov decadeofadversarialexamplesasurveyonthenatureandunderstandingofneuralnetworknonrobustness AT eelimonova decadeofadversarialexamplesasurveyonthenatureandunderstandingofneuralnetworknonrobustness AT vvarlazarov decadeofadversarialexamplesasurveyonthenatureandunderstandingofneuralnetworknonrobustness |