Geodesic Flow Matching for Denoising High-Dimensional Structured Representations

2026-05-29

AI Summary

Existing denoising methods overlook the geometric constraints imposed by the toroidal manifold on which spatial semantic pointers (SSPs) reside, leading to interpolation that disrupts their phase and magnitude structure and thereby degrading decoding accuracy. This work proposes a geodesic flow matching approach that, for the first time, integrates Riemannian transport dynamics into the denoising of high-dimensional structured representations, ensuring that the generated flow remains strictly confined to the SSP toroidal manifold. By doing so, it achieves geometry-aware neuro-symbolic cleanup. The method effectively circumvents the manifold distortion inherent in Euclidean assumptions, yielding a 72% reduction in tracking error and a 40% improvement in neural efficiency on path integration tasks, significantly outperforming current baselines.

Abstract

Vector Symbolic Algebras (VSAs) enable robust neurosymbolic reasoning by encoding symbolic information into high-dimensional distributed representations. For continuous domains, Spatial Semantic Pointers (SSPs) extend this framework by mapping variables onto continuous toroidal manifolds. However, standard approaches like Flow Matching assume a flat Euclidean geometry, which fails to account for the geometric constraints imposed on valid SSP states. We demonstrate that this assumption fails for SSPs: Euclidean linear interpolants ``cut through" the manifold's interior, destroying the phase and magnitude structure required for accurate decoding. To resolve this, we employ Geodesic Flow Matching, adapting Riemannian transport dynamics to strictly restrict the denoising flow to the SSP toroidal manifold. We validate this approach in a Spiking Neural SLAM system, showing that manifold-aware cleanup stabilizes path integration against drift. The method achieves a 72\% reduction in tracking error and enables a 40\% increase in neural efficiency compared to competitive baselines. Code is available at https://github.com/kremHabashy/CleanupSSP .

Problem

• Vector Symbolic Algebras
• Spatial Semantic Pointers
• Geodesic Flow Matching
• Manifold Geometry
• Denoising

Innovation

• Geodesic Flow Matching
• Spatial Semantic Pointers
• Riemannian manifold
• Vector Symbolic Algebras
• denoising