“…Moreover, the complexity of explicitly computing the matrix entries is bounded from below by the storage complexity. This bottleneck is greatly exacerbated in cases where IGA-based models are utilized within parametric studies, such as shape optimization [22,42,52,62], uncertainty quantification (UQ) for stochastic geometry deformations [24,63], or studies utilizing shape morphing techniques [64]. These parametric studies demand that multiple, often numerous, geometry configurations must be explored until an optimal shape or a statistical quantity of interest (QoI) can be estimated to sufficient accuracy.…”