Backdoor attacks in neural networks
As artificial intelligence becomes increasingly integrated in our daily lives, neural networks can be found in applications of deep learning in a multitude of critical domains, encompassing facial recognition, autonomous vehicular systems, and more. This pervasive integration, while transformative,...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project (FYP) |
| Language: | English |
| Published: |
Nanyang Technological University
2024
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/175146 |
| _version_ | 1824454280035172352 |
|---|---|
| author | Liew, Sher Yun |
| author2 | Zhang Tianwei |
| author_facet | Zhang Tianwei Liew, Sher Yun |
| author_sort | Liew, Sher Yun |
| collection | NTU |
| description | As artificial intelligence becomes increasingly integrated in our daily lives, neural networks can be found in applications of deep learning in a multitude of critical domains, encompassing facial recognition, autonomous vehicular systems, and more. This pervasive integration, while transformative, has brought about a pressing concern: the potential for disastrous consequences arising from malicious backdoor attacks in neural networks. To determine the effects and limitations of these attacks, this project aims to conduct a comprehensive examination of 2 previously proposed backdoor attack strategies, namely Blended and Blind backdoors, along with 2 previously proposed backdoor defence mechanisms, namely Neural Cleanse and Spectral Signatures. An exhaustive review of pertinent research literature was performed. Additionally, experiments were carried out to test the effectiveness of these strategies. |
| first_indexed | 2025-02-19T03:19:48Z |
| format | Final Year Project (FYP) |
| id | ntu-10356/175146 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2025-02-19T03:19:48Z |
| publishDate | 2024 |
| publisher | Nanyang Technological University |
| record_format | dspace |
| spelling | ntu-10356/1751462024-04-26T15:41:06Z Backdoor attacks in neural networks Liew, Sher Yun Zhang Tianwei School of Computer Science and Engineering tianwei.zhang@ntu.edu.sg Computer and Information Science Cyber security As artificial intelligence becomes increasingly integrated in our daily lives, neural networks can be found in applications of deep learning in a multitude of critical domains, encompassing facial recognition, autonomous vehicular systems, and more. This pervasive integration, while transformative, has brought about a pressing concern: the potential for disastrous consequences arising from malicious backdoor attacks in neural networks. To determine the effects and limitations of these attacks, this project aims to conduct a comprehensive examination of 2 previously proposed backdoor attack strategies, namely Blended and Blind backdoors, along with 2 previously proposed backdoor defence mechanisms, namely Neural Cleanse and Spectral Signatures. An exhaustive review of pertinent research literature was performed. Additionally, experiments were carried out to test the effectiveness of these strategies. Bachelor's degree 2024-04-22T05:47:49Z 2024-04-22T05:47:49Z 2024 Final Year Project (FYP) Liew, S. Y. (2024). Backdoor attacks in neural networks. Final Year Project (FYP), Nanyang Technological University, Singapore. https://hdl.handle.net/10356/175146 https://hdl.handle.net/10356/175146 en application/pdf Nanyang Technological University |
| spellingShingle | Computer and Information Science Cyber security Liew, Sher Yun Backdoor attacks in neural networks |
| title | Backdoor attacks in neural networks |
| title_full | Backdoor attacks in neural networks |
| title_fullStr | Backdoor attacks in neural networks |
| title_full_unstemmed | Backdoor attacks in neural networks |
| title_short | Backdoor attacks in neural networks |
| title_sort | backdoor attacks in neural networks |
| topic | Computer and Information Science Cyber security |
| url | https://hdl.handle.net/10356/175146 |
| work_keys_str_mv | AT liewsheryun backdoorattacksinneuralnetworks |