A Stopping Criterion for the Training Process of the Specific Signal Generator

Abstract

Mathematical description of a complex signal is very important in engineering but nearly impossible in many occasions. The emergence of the Generative Adversarial Network (GAN) shows the possibility to train a single neural network to be a Specific Signal Generator (SSG), which is only controlled by a random vector with several elements. However, there is no explicit criterion for the GAN training process to stop, and in real applications the training always stops after a certain big iteration. In this paper, a serious issue was discussed during the process to use GAN as a SSG. And, an explicit criterion for the GAN as a SSG to stop the training process were proposed. Several experiments were carried out to illustrate the issues mentioned above and the effectiveness of the stopping criterion proposed in this paper.

Improving the quality of generative models through Smirnov transformation

“…Furthermore, there is no consensus on the deterministic stopping criteria during GAN training that produce highquality synthetic data. Several works have already started to benchmark GANs to measure their evolution with respect to the quality of the generated data (for instance, [16,14,29]) and only a few of them have proposed a detention criterium not based on the visual inspection of the synthetic data (e.g., [5] and [17]). In fact, the former of these two papers proposes a stopping criterion that is closely related to the signal they generate, so that its direct application in other domains outside signal processing seems not very feasible.…”