Exploiting the relationship between Kendall’s rank correlation and cosine similarity for attribution protection

• المؤلفون الرئيسيون: Wang, Fan, Kong, Adams Wai Kin
• مؤلفون آخرون: School of Computer Science and Engineering
• التنسيق: Conference or Workshop Item
• اللغة: English
• منشور في: 2022
• الموضوعات: Computer Science - Learning; Computer Science - Artificial Intelligence; Artificial Intelligence; Neural Networks
• الوصول للمادة أونلاين: https://hdl.handle.net/10356/161935; https://nips.cc/
• المؤسسة: Nanyang Technological University

Model attributions are important in deep neural networks as they aid practitioners in understanding the models, but recent studies reveal that attributions can be easily perturbed by adding imperceptible noise to the input. The non-differentiable Kendall's rank correlation is a key performan...