に平均化+flatten Secret Key X (fixed) Weight w Sigmoid 出⼒ 透かし 埋め込みロス 1 0 1 1 E 0 E R E = E 0 + λE R Parameter regularizer として実装可能 (cf. weight decay) タスクのロスおよび 埋め込みロスを同時に最⼩化
classes) - 50,000 images for training - 10,000 images for test ネットワークアー キテクチャおよび パラメータ - WideResNet [4] (N = 1, k = 4) - SGD with Nesterov momentum - cross-entropy loss - the initial learning rate = 0.1 - weight decay = 5.0 x10-4 - momentum = 0.9 - minibatch size = 64 - λ = 0.01 電⼦透かし 256 bit (T = 256) 埋め込み対象 conv2 group [4] S. Zagoruyko and N. Komodakis. Wide residual networks. In Proc. of ECCV, 2016. conv1 conv2 group conv3 group conv4 group arg-pool fc M= 36864(3 x 3 x 64 x 64 )
CIFAR-10) ! どちらのケースでもfine-tuningで透かしは消えない テストエラーも埋め込みなし (8.04%) と同等 Note: Caltech-101 dataset were resized to 32 x 32 for compatibility with the CIFAR-10 dataset though their original sizes is roughly 300 x 200. 埋め込みロス before after
6], parameter pruning [5, 6] Robustness: model compression [5] S. Han, H. Mao, and W. J. Dally. Deep compression: Compressing deep neural networks with pruning, trained quantization and huffman coding. In Proc. of ICLR, 2016. [6] S. Han, J. Pool, J. Tran, and W. J. Dally. Learning both weights and connections for efficient neural networks. In Proc. of NIPS, 2015.
minima, and all local minima are almost optimal [8, 9]. Why Did Our Approach Work So Well? [7] A. Choromanska et al. The loss surfaces of multilayer networks. In Proc. of AISTATS, 2015. [8] Y. Dauphin et al. Identifying and attacking the saddle point problem in high-dimensional non-convex optimization. In Proc. of NIPS, 2014. Loss Parameter space Standard SGD
minima, and all local minima are almost optimal [8, 9]. ! Our embedding regularizer guides model parameters toward a local minima, which has the desired watermark. ! Let us assume that we want to embed the watermark “11”… Why Did Our Approach Work So Well? [7] A. Choromanska et al. The loss surfaces of multilayer networks. In Proc. of AISTATS, 2015. [8] Y. Dauphin et al. Identifying and attacking the saddle point problem in high-dimensional non-convex optimization. In Proc. of NIPS, 2014. Loss Parameter space 00 01 10 11 Detected watermark Standard SGD SGD with Embedding Loss
more details, please refer to… Y. Uchida, Y. Nagai, S. Sakazawa, and S. Satoh, “Embedding Watermarks into Deep Neural Networks,” in Proc. of International Conference on Multimedia Retrieval 2017.