SOURCE-LINKED INTELLIGENCE

Derivative-Free Structured Updates for Muon

arXiv · AI, language, vision and robotics · article · Sep 15, 2026 · UTC

Muon updates matrix-valued neural-network parameters by orthogonalizing a gradient-based momentum matrix. Its reliance on derivatives limits its use when gradients are unavailable or unreliable. We develop a derivative-free framework that constructs Muon-style updates from structured finite differences. Four variants are considered: full entrywise recovery, random low-rank surrogates, basis-aligned rank-one probing, and direct structured search. Exhaustive basis-aligned probing is equivalent, up to positive scaling before ideal polar orthogonalization, to coordinate finite differences. Matrix-

Read original source ↗ Open in workspace

recordType

paper

region

Global