Orth-Dion: Eliminating Geometric Mismatch in Distributed Low-Rank Spectral Optimization

• URL: http://arxiv.org/abs/2605.16341v1
• Date: Thu, 07 May 2026 23:37:37 GMT
• Title: Orth-Dion: Eliminating Geometric Mismatch in Distributed Low-Rank Spectral Optimization
• Abstract summary: Low-rank gradient compression reduces communication in distributed training by representing updates with rank-$r$ factors.<n>We propose Orth-Dion, which replaces column normalization with QR orthogonalization of the right factor.<n>Experiments on large-scale language model pre-training validate the predicted $sqrtr$ scaling and show that Orth-Dion closes the convergence gap to Muon at Dion's communication cost.
• Score: 12.526290684448425
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Low-rank gradient compression reduces communication in distributed training by representing updates with rank-$r$ factors. Dion is a recent method that approximates Muon, a spectral optimizer that orthogonalizes momentum, using one step of power iteration followed by column normalization (rescaling each column of the right factor to unit length). This makes it compatible with fully sharded data parallel training, but it converges more slowly than full-rank spectral methods. We show that this gap is geometric: column normalization does not yield the rank-$r$ polar factor that Muon implicitly targets, so the resulting direction violates the dual-norm constraint of the low-rank spectral geometry, and the rate picks up an extra factor of $\sqrt{r}$ even though the low-rank approximation of the gradient itself is accurate. The same mismatch enters the smoothness term and the error-feedback recursion in the analysis, which has a knock-on effect on empirical performance. We propose Orth-Dion, which replaces column normalization with QR orthogonalization of the right factor. Under non-Euclidean smoothness, with $L_r$ the curvature constant along rank-$r$ directions, Orth-Dion attains rate $O(\sqrt{L_r/T})$, matching exact spectral methods at the same per-step communication cost as Dion. The proof removes the bounded-drift assumption common in prior error-feedback analyses via a self-consistent fixed-point argument, and uses a time-averaged contraction that only requires the error sequence to contract on average rather than at every step. Experiments on large-scale language model pre-training validate the predicted $\sqrt{r}$ scaling and show that Orth-Dion closes the convergence gap to Muon at Dion's communication cost.

Related papers

• Beyond the Matrix Sign: Quadratic Spectral Descent [6.589542614722917]
  We show that curvature can change both the singular values and the singular directions of the optimal update.<n>Experiments on GPT pre-training show that QSD consistently improves validation loss over Muon and recent Muon variants.
  arXiv  Detail & Related papers  (2026-09-07T15:07:24Z)
• Demixing Sparse Signals from Nonlinear Observations using Generalized Non-convex Regularization [0.0]
  We consider the recovery of a pair of sparse vectors from a limited number of nonlinear observations of their superposition.<n>We propose a regularization-based framework combining a Huberized data fidelity with generalized folded-concave penalties.<n>We derive estimation error bounds of order $sqrtslog(n)/m$ that hold at emphevery localized stationary point.
  arXiv  Detail & Related papers  (2026-07-12T07:29:39Z)
• The Spectral Edge Thesis: A Mathematical Framework for Intra-Signal Phase Transitions in Neural Network Training [0.0]
  We show that phase transitions in neural network training are controlled by the spectral gap of the rolling-window Gram matrix of parameter updates.<n>Adiabatic parameter $mathcalA = |G|_F / (, g2)$ controls circuit stability: $mathcalA ll 1$ (plateau), $mathcalA sim 1$ (phase transition), $mathcalA gg 1$ (forgetting)
  arXiv  Detail & Related papers  (2026-03-30T20:10:22Z)
• Stability and Generalization of Push-Sum Based Decentralized Optimization over Directed Graphs [55.77845440440496]
  Push-based decentralized communication enables optimization over communication networks, where information exchange may be asymmetric.<n>We develop a unified uniform-stability framework for the Gradient Push (SGP) algorithm.<n>A key technical ingredient is an imbalance-aware generalization bound through two quantities.
  arXiv  Detail & Related papers  (2026-02-24T05:32:03Z)
• Optimal Unconstrained Self-Distillation in Ridge Regression: Strict Improvements, Precise Asymptotics, and One-Shot Tuning [61.07540493350384]
  Self-distillation (SD) is the process of retraining a student on a mixture of ground-truth and the teacher's own predictions.<n>We show that for any prediction risk, the optimally mixed student improves upon the ridge teacher for every regularization level.<n>We propose a consistent one-shot tuning method to estimate $star$ without grid search, sample splitting, or refitting.
  arXiv  Detail & Related papers  (2026-02-19T17:21:15Z)
• OLion: Approaching the Hadamard Ideal by Intersecting Spectral and $\ell_{\infty}$ Implicit Biases [29.60546958677364]
  nameA combines spectral control from update directions with coordinate control from sign updates.<n>We prove convergence under a mild, empirically verified diagonal-isotropy assumption.<n>nameA matches or outperforms AdamW and Muon under comparable tuning while using only momentum-level state.
  arXiv  Detail & Related papers  (2026-02-01T08:59:45Z)
• Avoiding Premature Collapse: Adaptive Annealing for Entropy-Regularized Structural Inference [1.7523718031184992]
  We identify a fundamental mechanism for this failure: textbfPremature Mode Collapse.<n>We propose textbfEfficient Piecewise Hybrid Adaptive Stability Control (EPH-ASC), an adaptive scheduling algorithm that monitors the stability of the inference process.
  arXiv  Detail & Related papers  (2026-01-30T14:47:18Z)
• Scale-Invariant Regret Matching and Online Learning with Optimal Convergence: Bridging Theory and Practice in Zero-Sum Games [60.871651115241406]
  A considerable chasm has been looming for decades between theory and practice in zero-sum game solving through first-order methods.<n>We propose a new scale-invariance and parameter-free variant of PRM$+$, which we call IREG-PRM$+$.<n>We show that it achieves $T-1/2$ best-iterate and $T-1$ optimal convergence guarantees, while also being on par with PRM$+$ on benchmark games.
  arXiv  Detail & Related papers  (2025-10-06T00:33:20Z)
• Convergence Rate Analysis of LION [54.28350823319057]
  LION converges iterations of $cal(sqrtdK-)$ measured by gradient Karush-Kuhn-T (sqrtdK-)$. We show that LION can achieve lower loss and higher performance compared to standard SGD.
  arXiv  Detail & Related papers  (2024-11-12T11:30:53Z)
• Adaptive, Doubly Optimal No-Regret Learning in Strongly Monotone and Exp-Concave Games with Gradient Feedback [75.29048190099523]
  Online gradient descent (OGD) is well known to be doubly optimal under strong convexity or monotonicity assumptions. In this paper, we design a fully adaptive OGD algorithm, textsfAdaOGD, that does not require a priori knowledge of these parameters.
  arXiv  Detail & Related papers  (2023-10-21T18:38:13Z)
• Sharper Convergence Guarantees for Asynchronous SGD for Distributed and Federated Learning [77.22019100456595]
  We show a training algorithm for distributed computation workers with varying communication frequency. In this work, we obtain a tighter convergence rate of $mathcalO!!!(sigma2-2_avg!! . We also show that the heterogeneity term in rate is affected by the average delay within each worker.
  arXiv  Detail & Related papers  (2022-06-16T17:10:57Z)
• Thinking Outside the Ball: Optimal Learning with Gradient Descent for Generalized Linear Stochastic Convex Optimization [37.177329562964765]
  We consider linear prediction with a convex Lipschitz loss, or more generally, convex optimization problems of generalized linear form. We show that in this setting, early iteration stopped Gradient Descent (GD), without any explicit regularization or projection, ensures excess error at most $epsilon$. But instead of uniform convergence in a norm ball, which we show can guarantee suboptimal learning using $Theta (1/epsilon4)$ samples, we rely on uniform convergence in a distribution-dependent ball.
  arXiv  Detail & Related papers  (2022-02-27T09:41:43Z)
• Optimal policy evaluation using kernel-based temporal difference methods [78.83926562536791]
  We use kernel Hilbert spaces for estimating the value function of an infinite-horizon discounted Markov reward process. We derive a non-asymptotic upper bound on the error with explicit dependence on the eigenvalues of the associated kernel operator. We prove minimax lower bounds over sub-classes of MRPs.
  arXiv  Detail & Related papers  (2021-09-24T14:48:20Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.