Leveraging imperfect restoration for data availability attack

• المؤلفون الرئيسيون: Huang, Yi, Styborski, Jeremy, Lyu, Mingzhi, Wang, Fan, Kong, Adams Wai Kin
• مؤلفون آخرون: Interdisciplinary Graduate School (IGS)
• التنسيق: Conference or Workshop Item
• اللغة: English
• منشور في: 2024
• الموضوعات: Computer and Information Science; Data availability attacks; Supervised learning; Self-supervised learning
• الوصول للمادة أونلاين: https://hdl.handle.net/10356/179131; https://eccv.ecva.net/virtual/2024/poster/1216
• المؤسسة: Nanyang Technological University

The abundance of online data is at risk of unauthorized usage in training deep learning models. To counter this, various Data Availability Attacks (DAAs) have been devised to make data unlearnable for such models by subtly perturbing the training data. However, existing attacks often excel against e...