Introduction

 Qutat Waveform is a simulation and data science framework designed to accelerate wave optics research. Along with a GPU-accelerated FDTD (Finite Difference Time Domain) simulator, it provides convenient features for creating and utilizing simulation datasets. Waveform focuses more on creating and using 'simulation datasets' rather than individual simulation results. It offers a modeling method to effectively express randomized structures that can form the input dataset, and SPDM (Simulation Process and Data Management) features to obtain simulation results from each of them, which can form the output dataset.

 On Waveform, machine learning models can be built using the dataset. Additionally, the prediction of simulation results, inverse designs, and optimizations can be conducted using these models. Although it contains a built-in wave optics solver, other user-defined solvers can also be integrated. Any simulation tool with a Python API can be connected and executed as a subprocess from the Waveform user interface. Therefore, many core-algorithm-only solvers can be integrated with several features of Waveform and easily scaled up to expand its functions, computational power, and user base.

Features