Variance-Reduced Manifold Sampling via Polynomial-Maximization Density Estimation

PMM-MASEM replaces k-nearest-neighbor plug-in density estimates in manifold sampling with a polynomial moment maximization estimator to reduce variance in resampling weights. Targets constrained simulation, motion planning, and probabilistic ML. Incremental improvement over MASEM; relevant to practitioners where density estimation noise degrades sampling quality on implicitly defined manifolds.