About this document

Detecting Inversion Attacks on DNN IP by Qe Singh is a document available to read on EtoBox.

This paper addresses the vulnerability of deep neural networks (DNNs) to generative model inversion attacks, which can bypass existing IP protection techniques like watermarking and fingerprinting. The authors propose a method called InverseDataInspector (IDI) to detect whether data has been inverted from victim models, thereby enabling the identification of IP infringement. Experimental results indicate that IDI effectively generalizes across various inversion methods and demonstrates high performance in d

Author

Qe Singh

Language

EN