• Abstract

In this paper, we first reinvestigate the convergence of the vanilla SGD method in the sense of $L^2$ under more general learning rates conditions and a more general convex assumption, which relieves the conditions on learning rates and does not need the problem to be strongly convex. Then, by taking advantage of the Lyapunov function technique, we present the convergence of the momentum SGD and Nesterov accelerated SGD methods for the convex and non-convex problem under $L$-smooth assumption that extends the bounded gradient limitation to a certain extent. The convergence of time averaged SGD was also analyzed.

• General Summary

• Abstract

We present a new category of physics-informed neural networks called physics informed variational embedding generative adversarial network (PI-VEGAN), that effectively tackles the forward, inverse, and mixed problems of stochastic differential equations. In these scenarios, the governing equations are known, but only a limited number of sensor measurements of the system parameters are available. We integrate the governing physical laws into PI-VEGAN with automatic differentiation, while introducing a variational encoder for approximating the latent variables of the actual distribution of the measurements. These latent variables are integrated into the generator to facilitate accurate learning of the characteristics of the stochastic partial equations. Our model consists of three components, namely the encoder, generator, and discriminator, each of which is updated alternatively employing the stochastic gradient descent algorithm. We evaluate the effectiveness of PI-VEGAN in addressing forward, inverse, and mixed problems that require the concurrent calculation of system parameters and solutions. Numerical results demonstrate that the proposed method achieves satisfactory stability and accuracy in comparison with the previous physics-informed generative adversarial network (PI-WGAN).

• General Summary

• Abstract

We present an innovative interpretation of Kalman filter (KF) combining the ideas of Schwarz domain decomposition (DD) and parallel in time (PinT) approaches. Thereafter we call it DD-KF. In contrast to standard DD approaches which are already incorporated in KF and other state estimation models, implementing a straightforward data parallelism inside the loop over time, DD-KF ab-initio partitions the whole model, including filter equations and dynamic model along both space and time directions/steps. As a consequence, we get local KFs reproducing the original filter at smaller dimensions on local domains. Also, sub problems could be solved in parallel. In order to enforce the matching of local solutions on overlapping regions, and then to achieve the same global solution of KF, local KFs are slightly modified by adding a correction term keeping track of contributions of adjacent subdomains to overlapping regions. Such a correction term balances localization errors along overlapping regions, acting as a regularization constraint on local solutions. Furthermore, such a localization excludes remote observations from each analyzed location improving the conditioning of the error covariance matrices. As dynamic model we consider shallow water equations which can be regarded a consistent tool to get a proof of concept of the reliability assessment of DD-KF in monitoring and forecasting of weather systems and ocean currents.

• General Summary

• Abstract

The problem of finding fixed points of nonexpansive mappings on Hadamard manifolds is considered in this paper. To solve this kind of problem, we propose a modified Riemannian Mann algorithm with inertial effect. Under the assumption of existence of fixed points of the nonexpansive mapping, the global convergence of the proposed algorithm is established. To show the efficiency of the proposed algorithm, numerical comparisons with some existing algorithms are reported.

• General Summary