To address the coupling problem, we propose a Cycle Self-Training (CST) framework for SSOD, which consists of two teachers T1 and T2, two students S1 and S2.
Instead of performing stylization frame by frame, only key frames in the original video are processed by a pre-trained deep neural network (DNN) on edge servers, while the rest of stylized intermediate frames are generated by our designed optical-flow-based frame interpolation algorithm on mobile phones.
In this work, we derive a new regularized risk, in which the regularizer can adaptively encourage the accuracy and robustness of the smoothed counterpart when training the base classifier.
As such, training our GAN architecture requires much fewer high-quality images with a small number of additional low-quality images.
We intuitively and empirically prove the rationality of our method in reducing the search space.
It becomes an open question whether escaping sharp minima can improve the generalization.
Our experiments show that different adversarial strengths, i. e., perturbation levels of adversarial examples, have different working zones to resist the attack.
Our DNN has 4. 1M parameters, which is only 6. 7% of AlexNet or 59% of GoogLeNet.
SSL can: (1) learn a compact structure from a bigger DNN to reduce computation cost; (2) obtain a hardware-friendly structured sparsity of DNN to efficiently accelerate the DNNs evaluation.
IBM TrueNorth chip uses digital spikes to perform neuromorphic computing and achieves ultrahigh execution parallelism and power efficiency.