Generative adversarial network based on chaotic time series

Abstract

Generative adversarial networks (GANs) are becoming increasingly important in the artificial construction of natural images and related functionalities, wherein two types of networks called generators and discriminators evolve through adversarial mechanisms. Using deep convolutional neural networks and related techniques, high-resolution and highly realistic scenes, human faces, etc. have been generated. GANs generally require large amounts of genuine training data sets, as well as vast amounts of pseudorandom numbers. In this study, we utilized chaotic time series generated experimentally by semiconductor lasers for the latent variables of a GAN, whereby the inherent nature of chaos could be reflected or transformed into the generated output data. We show that the similarity in proximity, which describes the robustness of the generated images with respect to minute changes in the input latent variables, is enhanced, while the versatility overall is not severely degraded. Furthermore, we demonstrate that the surrogate chaos time series eliminates the signature of the generated images that is originally observed corresponding to the negative autocorrelation inherent in the chaos sequence. We also address the effects of utilizing chaotic time series to retrieve images from the trained generator.

Natural Exponential and Three‐Dimensional Chaotic System

“…[ 43 ] Many attempts have been made on natural exponential functions and methods in improving the performance of traditional chaotic models applied to weak signal detection, [ 33,43 ] ecosystem experiment, [ 44 ] quantum computation, [ 45 ] astronomical observation, [ 46 ] geological prediction, [ 47 ] and network generation. [ 48 ] For instance, the hidden attractor discovery with an exponential nonlinear term [ 49 ] shows rich dynamic behaviors, and the out‐of‐time‐ordered correlation models (OTOCs) proportional to a natural exponential function is also presented to quantify weak quantum chaos [ 50 ] and many‐body system, [ 51 ] higher dimensional characterizations of a chaotic system, especially for the natural exponential function based models, are rarely studied and revealed. [ 52 ] If three or higher dimensional chaotic system is expanded or coupled with 2D classical models, better performance of the chaotic system may be shown.…”

A framework for data-driven solution and parameter estimation of PDEs using conditional generative adversarial networks

“…The training phase is complete when the discriminator cannot distinguish between real data belonging to the generator's training set and data produced by the generator. The concept of deep convolutional neural networks (CNN) has been employed for both the generator and the discriminator of GAN to capture highly heterogeneous spatial features [46][47][48] and their evolution in time [49,50]. Additionally, an augmentation of discrete labeled data has been used to control the generator's output, termed conditional GAN or cGAN [51][52][53][54].…”

Generation of Schubert polynomial series via nanometre-scale photoisomerization in photochromic single crystal and double-probe optical near-field measurements

“…Irregular time series play critical roles in information and communication technology today, including in secure information transfer 1,2 , Monte Carlo simulations 3 , and machine learning 4,5 . Physical processes in nature are interesting resources for providing irregular time series, including deterministic dynamics such as chaos 6 , rather than only truly random sequences such as those caused by single photons 7 .…”

“…Physical processes in nature are interesting resources for providing irregular time series, including deterministic dynamics such as chaos 6 , rather than only truly random sequences such as those caused by single photons 7 . Indeed, chaotic lasers enable interesting functionalities ranging from ultrafast random number generation 8 and photonic reservoir computing 9 to decision-making, reinforcement learning 10 , and artificial data generation 5 . In the case of chaotic time series, a minute initial difference results in significantly different series, while the series share common attributes specified by the dynamics therein.…”