paper-with-me

Papers

Dimension-Free Saddle-Point Escape in Muon

2026-05-10 · Yanlin Long, Yufei Gu, Zeke Xie arxiv

Modern Large Language Model (LLM) training is fundamentally bottlenecked by pathologically flat saddle points in extreme high-dimensional landscapes. Motivated by this challenge, we analyze the saddle-point escape dynamics of the emerging Muon optimizer, demonstrating its resilience against the $\mathcal{O}(D)$ dimensional curse that severely traps element-wise adaptive optimizers like AdamW. By extending generalized matrix perturbation theory, we develop a theoretical framework to capture Muon's non-equilibrium optimization trajectories. This theoretical machinery mathematically proves that Muon elegantly bypasses the dimensional curse via a non-linear spectral shaping mechanism. By leveraging resolvent functional calculus and macroscopic Cauchy contour integration, we avoid isotropic noise assumptions and Tracy-Widom edge singularities. We establish that structural incoherence securely shields the trajectory from orthogonal drift, enabling a dimension-free saddle-point escape, and triggering a deterministic $\mathcal{O}(1)$ discrete ballistic ejection under sufficient spectral gap. Consequently, we provide an algebraically dimension-free escape bound for Muon, formalizing the underlying mechanics of its non-convex optimization dynamics.

📄 PDF Abstract BibTeX arXiv:2605.09331

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

How to Escape Saddle Points Efficiently

2017-03-02 · ICML 2017 8 · Chi Jin, Rong Ge, Praneeth Netrapalli, Sham M. Kakade 외

This paper shows that a perturbed form of gradient descent converges to a second-order stationary point in a number iterations which depends only poly-logarithmically on dimension (i.e., it is almost "dimension-free"). T…

Notes on Worst-case Inefficiency of Gradient Descent Even in R^2

2020-08-17 · Shiliang Zuo

Gradient descent is a popular algorithm in optimization, and its performance in convex settings is mostly well understood. In non-convex settings, it has been shown that gradient descent is able to escape saddle points a…

Adaptive Stochastic Gradient Langevin Dynamics: Taming Convergence and Saddle Point Escape Time

2018-05-23 · Hejian Sang, Jia Liu

In this paper, we propose a new adaptive stochastic gradient Langevin dynamics (ASGLD) algorithmic framework and its two specialized versions, namely adaptive stochastic gradient (ASG) and adaptive gradient Langevin dyna…

On the saddle point problem for non-convex optimization

2014-05-19 · Razvan Pascanu, Yann N. Dauphin, Surya Ganguli, Yoshua Bengio

A central challenge to many fields of science and engineering involves minimizing non-convex error functions over continuous, high dimensional spaces. Gradient descent or quasi-Newton methods are almost ubiquitously used…

Escaping Saddle Points with Stochastically Controlled Stochastic Gradient Methods

2021-03-07 · Guannan Liang, Qianqian Tong, Chunjiang Zhu, Jinbo Bi

Stochastically controlled stochastic gradient (SCSG) methods have been proved to converge efficiently to first-order stationary points which, however, can be saddle points in nonconvex optimization. It has been observed …