Spectral integrated neural networks (SINNs) for solving forward and inverse dynamic problems.

This study introduces an innovative neural network framework named spectral integrated neural networks (SINNs) to address both forward and inverse dynamic problems in three-dimensional space. In the SINNs, the spectral integration technique is utilized for temporal discretization, followed by the application of a fully connected neural network to solve the resulting partial differential equations in the spatial domain. Furthermore, the polynomial basis functions are employed to expand the unknown function, with the goal of improving the performance of SINNs in tackling inverse problems. The performance of the developed framework is evaluated through several dynamic benchmark examples encompassing linear and nonlinear heat conduction problems, linear and nonlinear wave propagation problems, inverse problem of heat conduction, and long-time heat conduction problem. The numerical results demonstrate that the SINNs can effectively and accurately solve forward and inverse problems involving heat conduction and wave propagation. Additionally, the SINNs provide precise and stable solutions for dynamic problems with extended time durations. Compared to commonly used physics-informed neural networks, the SINNs exhibit superior performance with enhanced convergence speed, computational accuracy, and efficiency.

Anderson critical metal phase in trivial states protected by average magnetic crystalline symmetry.

Transitions between distinct obstructed atomic insulators (OAIs) protected by crystalline symmetries, where electrons form molecular orbitals centering away from the atom positions, must go through an intermediate metallic phase. In this work, we find that the intermediate metals will become a scale-invariant critical metal phase (CMP) under certain types of quenched disorder that respect the magnetic crystalline symmetries on average. We explicitly construct models respecting average C2zT, m, and C4zT and show their scale-invariance under chemical potential disorder by the finite-size scaling method. Conventional theories, such as weak anti-localization and topological phase transition, cannot explain the underlying mechanism. A quantitative mapping between lattice and network models shows that the CMP can be understood through a semi-classical percolation problem. Ultimately, we systematically classify all the OAI transitions protected by (magnetic) groups P m , P 2 ' , P 4 ' , and P 6 ' with and without spin-orbit coupling, most of which can support CMP.

A speculative extension of the differential operator definition to fractal via the fundamental solution.

This paper makes a speculative extension of the fundamental solution of the standard integer-order differential operators to fractal. Then, the fractal fundamental solution is used via the implicit calculus equation modeling approach to define differential operators on fractal for modeling complex mechanical behaviors of fractal materials. By employing the singular boundary method, a recent boundary discretization technique with the fundamental solution, this study also makes numerical simulation of fractal Laplace problems of multiply-connected and composite material. Results show the validity and rationality of the conjectured definition of Laplace operator on fractal. Furthermore, the fractional and the fractal Laplace operators are also compared in our numerical experiments.