English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

Survey of Optimizers: A 2025-2026 Review of Neural Network Optimization Methods

Forum topic · 小凯 · 2026-09-01

Summary

A survey paper (arXiv:2608.28557) by Ruoran Xu reorganizes the neural network optimizer landscape of 2025-2026. The author argues that modern optimization research can no longer be described as a simple succession of Adam variants: the design space has expanded from coordinate-wise updates to matrix- and layer-level methods, from fixed training horizons to time-based policies, and from mathematical update rules to state representations that must survive sharding and low-precision computation. The survey organizes recent work along four largely independent axes: temporal estimation, update geometry, horizon management, and representation and systems. It covers Muon's spectral normalization, Shampoo and SOAP's historical matrix statistics, adaptive and hybrid matrix methods, memory-efficient optimizers, schedule-free training, small-batch corrections, and quantized optimizer states. The core empirical conclusion is deliberately non-triumphal: matrix-aware methods represent genuine progress, but there is no context-free replacement for AdamW. Rankings vary with model scale, data-to-parameter ratio, batch size, schedule, parameter partitioning, tuning budget, and the target metric (tokens, FLOPs, wall-clock time, or memory).

Paper Overview

  • Field: Machine Learning
  • Author: Ruoran Xu
  • Published: 2026-08-28
  • arXiv: 2608.28557
  • Summary

    Neural-network optimization in 2025-2026 is no longer well described as a succession of new Adam variants. The design space has expanded from coordinates to matrices and layers, from fixed training horizons to policies over time, and from mathematical update rules to state representations that must survive sharding and low-precision computation.

    This survey organizes recent optimizers and training optimization methods along four largely independent axes:

    1. Temporal estimation — how gradients and statistics are estimated over time 2. Update geometry — the shape and structure of parameter updates (coordinate-wise vs. matrix-aware) 3. Horizon management — fixed training schedules vs. schedule-free policies 4. Representation and systems — optimizer state design under sharding and low-precision constraints

    Topics Covered

  • Spectral normalization in Muon
  • Historical matrix statistics in Shampoo and SOAP
  • Adaptive and hybrid matrix methods
  • Memory-efficient optimizers
  • Schedule-free training
  • Small-batch corrections
  • Quantized optimizer states
  • Key Takeaway

    The core empirical conclusion is deliberately non-triumphal: matrix-aware methods represent real progress, but there is no context-free replacement for AdamW. Rankings change with model scale, data-to-parameter ratio, batch size, schedule, parameter partitioning, tuning budget, and whether the target metric is tokens, FLOPs, wall-clock time, or memory.

    The practical consequences are a compositional view of optimizer design and stricter protocols for evaluating optimizer claims.

    Links

  • Paper: https://arxiv.org/abs/2608.28557
---

*Auto-collected on 2026-09-01.*

Tags

#optimizers#machine-learning#survey#adamw#muon#shampoo#training-efficiency#arxiv

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178634339