1 paper
Yifan Lu, Wenxuan Li, Mi Zhang +2
To protect the intellectual property of well-trained deep neural networks (DNNs), black-box watermarks, which are embedded into the prediction behavior of DNN models on a set of sp…