Simplex-FEM Networks (SiFEN): Learning A Triangulated Function Approximator
This work provides a compact, interpretable, and theoretically grounded alternative to dense MLPs and edge-spline networks for researchers and practitioners in machine learning and computational modeling.
The authors tackled the problem of function approximation by introducing Simplex-FEM Networks (SiFEN), a learned piecewise-polynomial predictor that achieves the classic FEM approximation rate M^(-m/d) and empirically matches or surpasses MLPs and KANs in performance while improving calibration and reducing inference latency.
We introduce Simplex-FEM Networks (SiFEN), a learned piecewise-polynomial predictor that represents f: R^d -> R^k as a globally C^r finite-element field on a learned simplicial mesh in an optionally warped input space. Each query activates exactly one simplex and at most d+1 basis functions via barycentric coordinates, yielding explicit locality, controllable smoothness, and cache-friendly sparsity. SiFEN pairs degree-m Bernstein-Bezier polynomials with a light invertible warp and trains end-to-end with shape regularization, semi-discrete OT coverage, and differentiable edge flips. Under standard shape-regularity and bi-Lipschitz warp assumptions, SiFEN achieves the classic FEM approximation rate M^(-m/d) with M mesh vertices. Empirically, on synthetic approximation tasks, tabular regression/classification, and as a drop-in head on compact CNNs, SiFEN matches or surpasses MLPs and KANs at matched parameter budgets, improves calibration (lower ECE/Brier), and reduces inference latency due to geometric locality. These properties make SiFEN a compact, interpretable, and theoretically grounded alternative to dense MLPs and edge-spline networks.