Search papers, labs, and topics across Lattice.
This paper investigates the implications of learning boundary terms in reflected diffusion processes on bounded domains, revealing that the learned score's boundary behavior can be misaligned with the forward process. By introducing the concept of conormal trace, the authors demonstrate that the no-flux condition can effectively determine the boundary term without introducing extra trainable parameters, thereby ensuring accurate representation of the true score under certain conditions. Their findings indicate that while hard reflection can maintain feasible samples, it may obscure errors in the learned boundary score, particularly under anisotropic diffusion scenarios.
Hard reflection can mask boundary-score errors in reflected diffusion, leading to misleading sample quality despite underlying inaccuracies.
We study score learning for reflected diffusion on bounded domains. Reflection keeps trajectories feasible but does not ensure that the learned score satisfies the boundary behavior implied by the forward process. With implicit score matching, integration by parts leaves a boundary term, and we show that it depends on one scalar at each boundary point: the diffusion- weighted normal component of the score, or conormal trace. The no-flux condition fixes this value while leaving the re- maining boundary components unrestricted; under anisotropic diffusion it generally differs from the ordinary normal score component. On hyperrectangles, our parametrization enforces the required trace without additional trainable parameters or a stochastic boundary estimator and, under regularity assump- tions, can represent the true score, whereas fixing an incorrect value creates an error that more data cannot remove. We ex- tend the construction to simplices and polygonal domains and identify reflection masking: hard reflection can keep samples feasible even when the learned trace is wrong, so post-reflection metrics may hide the error. Experiments show the clearest separation with less frequent reflection, anisotropic diffusion, and mass near intersections of constraints; under full reflection, final sample placement improves inconsistently, illustrating how hard repair can mask boundary-score errors and decouple score accuracy from downstream generation quality.
