“…However, most work to date has primarily focused on establishing the feasibility of carrying out nonlinear wind analysis and is based on direct integration approaches developed in seismic engineering. While recent advancements have occurred in this area-for example, the suite of methods that combine direct stochastic simulation with dynamic shakedown (Chuang and Spence 2019, the approaches based on reduced order models (Wu 2013;Wu and Kareem 2015;Zhao et al 2019;Li et al 2021;Li and Spence 2022a,b,c), and the methods that leverage machine learning (Li and Spence 2022a;Preetha Hareendran et al 2022)-much work remains to solve the problem of rapidly evaluating the nonlinear response of wind-excited structural systems in ways that are both robust to the complexity of the computational models that describe the nonlinear response of systems and compatible with general purpose uncertainty propagation schemes. Areas with promises in this respect are those related to metamodeling/surrogate modeling, reduced-order modeling, methods for leveraging massive parallelization through GPUs and supercomputing, and methods that leverage artificial intelligence (e.g., physics-informed and/or data-driven machine learning).…”