Spline Based

Spline-based methods are increasingly used to model complex, often irregularly sampled data, aiming for both accuracy and interpretability. Current research focuses on developing novel spline-based neural network architectures, Bayesian approaches for uncertainty quantification, and efficient algorithms for detecting and tracking splines in high-density data, such as in microscopy images or for modeling physical systems. These advancements are improving the analysis of diverse datasets across fields like biomedicine (e.g., spinal curvature estimation, nematode tracking), materials science (e.g., interatomic potentials), and time series analysis, enabling more robust and insightful modeling of complex phenomena.