# Low-rank Orthogonalization for Large-scale Matrix Optimization with Applications to Foundation Model Training

arXiv: https://arxiv.org/abs/2509.11983 · HTML: https://mlsanity.com/paper/2509.11983/

## Key facts

- Cited by 18 papers in the map (data through 14 Oct 2026).
- Builds on 27 of its 56 references that are in the map.
- First posted on arXiv 15 Sep 2025.

## Builds on (most cited first, 27 of 27)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [GaLore: Memory-Efficient LLM Training by Gradient Low-Rank Projection](https://mlsanity.com/paper/2403.03507/) (2024)
- [Muon is Scalable for LLM Training](https://mlsanity.com/paper/2502.16982/) (2025)
- [Shampoo: Preconditioned Stochastic Tensor Optimization](https://mlsanity.com/paper/1802.09568/) (2018)
- [Fine-Tuning Language Models with Just Forward Passes](https://mlsanity.com/paper/2305.17333/) (2023)
- [Old Optimizer, New Norm: An Anthology](https://mlsanity.com/paper/2409.20325/) (2024)
- [Training Deep Learning Models with Norm-Constrained LMOs](https://mlsanity.com/paper/2502.07529/) (2025)
- [Lower Bounds for Finding Stationary Points I](https://mlsanity.com/paper/1710.11606/) (2017)
- [Flora: Low-Rank Adapters Are Secretly Gradient Compressors](https://mlsanity.com/paper/2402.03293/) (2024)
- [Understanding Gradient Orthogonalization for Deep Learning via Non-Euclidean Trust-Region Optimization](https://mlsanity.com/paper/2503.12645/) (2025)
- [Muon Optimizes Under Spectral Norm Constraints](https://mlsanity.com/paper/2506.15054/) (2025)
- [A Note on the Convergence of Muon](https://mlsanity.com/paper/2502.02900/) (2025)
- [Practical Efficiency of Muon for Pretraining](https://mlsanity.com/paper/2505.02222/) (2025)
- [Gluon: Making Muon & Scion Great Again! (Bridging Theory and Practice of LMO-based Optimizers for LLMs)](https://mlsanity.com/paper/2505.13416/) (2025)
- [Dion: Distributed Orthonormalized Updates](https://mlsanity.com/paper/2504.05295/) (2025)
- [AdaMuon: Adaptive Muon Optimizer](https://mlsanity.com/paper/2507.11005/) (2025)
- [ASGO: Adaptive Structured Gradient Optimization](https://mlsanity.com/paper/2503.20762/) (2025)
- [PolarGrad: A Class of Matrix-Gradient Optimizers from a Unifying Preconditioning Perspective](https://mlsanity.com/paper/2505.21799/) (2025)
- [Lions and Muons: Optimization via Stochastic Frank-Wolfe under Heavy-Tailed Noise](https://mlsanity.com/paper/2506.04192/) (2025)
- [Structured Preconditioners in Adaptive Optimization: A Unified Analysis](https://mlsanity.com/paper/2503.10537/) (2025)
- [COSMOS: A Hybrid Adaptive Optimizer for Memory-Efficient Training of LLMs](https://mlsanity.com/paper/2502.17410/) (2025)
- [Orthogonalising gradients to speed up neural network optimisation](https://mlsanity.com/paper/2202.07052/) (2022)
- [Memory-Efficient LLM Pretraining via Minimalist Optimizer Design](https://mlsanity.com/paper/2506.16659/) (2025)
- [LoRA meets Riemannion: Muon Optimizer for Parametrization-independent Low-Rank Adapters](https://mlsanity.com/paper/2507.12142/) (2025)
- [Complexity of normalized stochastic first-order methods with momentum under heavy-tailed noise](https://mlsanity.com/paper/2506.11214/) (2025)
- [A Preconditioned Riemannian Gradient Descent Algorithm for Low-Rank Matrix Recovery](https://mlsanity.com/paper/2305.02543/) (2023)

## Built on by (most cited first, 18 of 18)

- [Isotropic Curvature Model for Understanding Deep Learning Optimization: Is Gradient Orthogonalization Optimal?](https://mlsanity.com/paper/2511.00674/) (2025)
- [The Newton-Muon Optimizer](https://mlsanity.com/paper/2604.01472/) (2026)
- [Insights on Muon from Simple Quadratics](https://mlsanity.com/paper/2602.11948/) (2026)
- [Powering Up Zeroth-Order Training via Subspace Gradient Orthogonalization](https://mlsanity.com/paper/2602.17155/) (2026)
- [Non-Euclidean SGD for Structured Optimization: Unified Analysis and Improved Rates](https://mlsanity.com/paper/2511.11466/) (2025)
- [Muon with Nesterov Momentum: Heavy-Tailed Noise and (Randomized) Inexact Polar Decomposition](https://mlsanity.com/paper/2605.06884/) (2026)
- [Rethinking Muon Beyond Pretraining: Spectral Failures and High-Pass Remedies for VLA and RLVR](https://mlsanity.com/paper/2605.19282/) (2026)
- [Why Muon Outperforms Adam: A Curvature Perspective](https://mlsanity.com/paper/2606.04662/) (2026)
- [Where Does Warm-Up Come From? Adaptive Scheduling for Norm-Constrained Optimizers](https://mlsanity.com/paper/2602.05813/) (2026)
- [DeMuon: A Decentralized Muon for Matrix Optimization over Graphs](https://mlsanity.com/paper/2510.01377/) (2025)
- [MiMuon: Mixed Muon Optimizer with Improved Generalization for Large Models](https://mlsanity.com/paper/2605.19619/) (2026)
- [Tight Complexity of Muon under Heavy-Tailed Noise](https://mlsanity.com/paper/2605.18528/) (2026)
- [Denoise First, Orthogonalize Later: Understanding Momentum in Muon via Spectral Filtering](https://mlsanity.com/paper/2606.03899/) (2026)
- [Subspace Control: Turning Constrained Model Steering into Controllable Spectral Optimization](https://mlsanity.com/paper/2604.04231/) (2026)
- [Position: Zeroth-Order Optimization in Deep Learning Is Underexplored, Not Underpowered](https://mlsanity.com/paper/2605.15622/) (2026)
- [Schattor: Schatten-family methods for deep learning optimization](https://mlsanity.com/paper/2606.15702/) (2026)
- [MALT: Lightweight Curvature-Aware Muon via Diagonal Preconditioning](https://mlsanity.com/paper/2608.05088/) (2026)
- [Approximate Muon with low-rank adapters](https://mlsanity.com/paper/2608.14492/) (2026)

## Abstract (arXiv)

Neural network (NN) training is inherently a large-scale matrix optimization problem, yet the matrix structure of NN parameters has long been overlooked. Recently, the optimizer Muon citepjordanmuon, which explicitly exploits this structure, has gained significant attention for its strong performance in foundation model training. A key component contributing to Muon's success is matrix orthogonalization. In this paper, we propose textitlow-rank orthogonalization, which performs orthogonalization by leveraging the low-rank nature of gradients during NN training. Building on this, we introduce low-rank matrix-signed gradient descent (MSGD) and a low-rank variant of Muon. %Numerical experiments demonstrate the superior performance of low-rank orthogonalization, with low-rank Muon achieving promising results in GPT-2 and LLaMA pretraining---surpassing the carefully tuned vanilla Muon on tasks with large model sizes. Numerical experiments demonstrate the advantages of low-rank orthogonalization: low-rank Muon generally matches or improves upon vanilla Muon on the GPT-2 and LLaMA pretraining tasks, with clearer improvements observed for relatively larger models. Theoretically, we establish the iteration complexity of low-rank MSGD for finding an approximate stationary solution, and the iteration complexity of low-rank Muon for finding an approximate stochastic stationary solution under heavy-tailed noise. The code to reproduce our numerical experiments is available at https://github.com/dengzhanwang/Low-rank-Muon.