Predicting Bulk Average Velocity with Rigid Vegetation in Open Channels Using Tree-Based Machine Learning: A Novel Approach Using Explainable Artificial Intelligence.

Predicting the bulk-average velocity (UB) in open channels with rigid vegetation is complicated due to the non-linear nature of the parameters. Despite their higher accuracy, existing regression models fail to highlight the feature importance or causality of the respective predictions. Therefore, we propose a method to predict UB and the friction factor in the surface layer (fS) using tree-based machine learning (ML) models (decision tree, extra tree, and XGBoost). Further, Shapley Additive exPlanation (SHAP) was used to interpret the ML predictions. The comparison emphasized that the XGBoost model is superior in predicting UB (R = 0.984) and fS (R = 0.92) relative to the existing regression models. SHAP revealed the underlying reasoning behind predictions, the dependence of predictions, and feature importance. Interestingly, SHAP adheres to what is generally observed in complex flow behavior, thus, improving trust in predictions.

Fabrication, characterization, and theoretical analysis of controlled disorder in the core of optical fibers.

We present results of experimental and theoretical studies of polarization-resolved light transmission through optical fiber with disorder generated in its germanium-doped core via UV radiation transmitted through a diffuser. In samples longer than a certain characteristic length, the power transmitted with preserved polarization is observed to be distributed over all forward-propagating modes, as evidenced by the Rayleigh negative exponential distribution of the near-field intensity at the output surface of the fiber. Furthermore, the transmitted power becomes also equally distributed over both polarizations. To describe the optical properties of the fibers with the experimentally induced disorder, a theoretical model based on coupled-mode theory is developed. The obtained analytical expression for the correlation function describing spatial properties of the disorder shows that it is highly anisotropic. Our calculations demonstrate that this experimentally controllable anisotropy can lead to suppression of the radiative leakage of the propagating modes, so that intermode coupling becomes the dominant scattering process. The obtained theoretical expressions for the polarization-resolved transmission fit very well with the experimental data, and the information extracted from the fit shows that radiative leakage is indeed small. The reported technique provides an easy way to fabricate different configurations of controlled disorder in optical fibers suitable for such applications as random fiber lasers.