AI RESEARCH

Explicit and Effectively Symmetric Schemes for Neural SDEs on Lie Groups

arXiv CS.LG

ArXi:2509.20599v2 Announce Type: replace Backpropagation through (neural) SDE solvers is traditionally approached in two ways: discretise-then-optimise, which offers accurate gradients but incurs prohibitive memory costs; and optimise-then-discretise, which achieves constant memory cost by solving an auxiliary backward SDE, but suffers from slower evaluation and gradient approximation errors.