Research Papers 论文研究 1d ago Updated 20h ago 更新于 20小时前 43

DeltaMomentum: A Key-Value based Anisotropic Momentum Update via Delta Rule DeltaMomentum:基于键值的各向异性动量更新方法

DeltaMomentum introduces a direction-aware momentum update rule that replaces the fixed-rate exponential moving average (EMA) with a key-value based delta rule, where each direction is forgotten at a rate proportional to its query frequency The gradient of a linear layer naturally decomposes into input (key) and output-side error (value), enabling anisotropic forgetting without matrix inversion DeltaAdamW reaches AdamW's validation loss in up to 46.39% fewer steps at 67M parameters and 22.12% fe 提出DeltaMomentum,一种基于Key-Value结构和Delta规则的各向异性动量更新方法,将方向感知能力直接嵌入动量更新规则 利用线性层梯度可分解为输入侧(key)和输出侧(value)的特性,通过Delta规则实现自适应遗忘速率 作为即插即用模块可替换任何优化器的动量缓冲,额外计算开销仅22.2%-25.0%,无需持久内存 在FineWeb-Edu预训练中,DeltaAdamW比AdamW减少46.39%步骤达到相同验证损失(67M模型),增益在1B模型Chinchilla最优预算下持续 理论证明该方法无需矩阵求逆即可实现输入侧曲率校正,且能更快清除陈旧方向

55
Hot 热度
72
Quality 质量
62
Impact 影响力

Analysis 深度分析

TL;DR

  • DeltaMomentum introduces a direction-aware momentum update rule that replaces the fixed-rate exponential moving average (EMA) with a key-value based delta rule, where each direction is forgotten at a rate proportional to its query frequency
  • The gradient of a linear layer naturally decomposes into input (key) and output-side error (value), enabling anisotropic forgetting without matrix inversion
  • DeltaAdamW reaches AdamW's validation loss in up to 46.39% fewer steps at 67M parameters and 22.12% fewer steps at 370M parameters on FineWeb-Edu pretraining
  • The method is a drop-in replacement for any optimizer's momentum buffer, transfers across widths under muP, and adds only 22.2%–25.0% extra compute relative to a gated-MLP block's linear cost with no persistent memory overhead
  • Gains generalize across architectures (ResNet-18, ViT-Tiny on CIFAR-10) and optimizers (SGD, Muon baseline), with training diagnostics confirming healthier input directions and better gradient tracking

Why It Matters

This work directly addresses a fundamental limitation of modern optimizers—their uniform forgetting rate for all gradient directions—which is increasingly costly as models scale. By embedding anisotropic awareness into the momentum update itself rather than wrapping external processing around it, DeltaMomentum offers a principled, low-overhead improvement that could become a standard upgrade for large-scale training pipelines.

Technical Details

  • Core mechanism: The gradient of a linear layer is decomposed into an input-side key and an output-side value. The momentum buffer is updated via the canonical delta rule, so frequently queried directions retain momentum longer while stale directions are cleared faster than EMA allows.
  • Theoretical guarantees: Proven to be a valid momentum, to apply input-side curvature correction without matrix inversion, and to outperform EMA in clearing stale directions under both fixed and drifting optima.
  • Efficiency: Extra compute is bounded between 22.2% and 25.0% of a gated-MLP block's linear cost, with no persistent memory requirement. The coefficient transfers across model widths under muP (mu-parameterization).
  • Empirical results: On FineWeb-Edu pretraining, DeltaAdamW achieves AdamW's validation loss in up to 46.39 ± 4.32% fewer steps at 67M and 22.12 ± 0.80% fewer steps at 370M (three seeds). Gains persist at 1B on a Chinchilla-optimal budget. A tuned Muon baseline underperforms DeltaAdamW at both language-model scales.
  • Generalization: Validated on SGD with ResNet-18 and ViT-Tiny on CIFAR-10. Training-time diagnostics confirm the predicted mechanism: better gradient tracking and healthier input directions.

Industry Insight

  • DeltaMomentum's drop-in compatibility means practitioners can adopt it immediately across existing optimizer stacks (AdamW, SGD, etc.) without architectural changes, making it a low-risk, high-reward upgrade for large-scale pretraining.
  • The observed speedup of up to ~46% at smaller scales suggests significant compute savings at production scale, where even marginal per-step efficiency gains translate to substantial cost reductions.
  • The anisotropic forgetting principle could extend beyond language models to vision, multimodal, and reinforcement learning domains where gradient direction frequency is similarly uneven, warranting broader empirical investigation.

TL;DR

  • 提出DeltaMomentum,一种基于Key-Value结构和Delta规则的各向异性动量更新方法,将方向感知能力直接嵌入动量更新规则
  • 利用线性层梯度可分解为输入侧(key)和输出侧(value)的特性,通过Delta规则实现自适应遗忘速率
  • 作为即插即用模块可替换任何优化器的动量缓冲,额外计算开销仅22.2%-25.0%,无需持久内存
  • 在FineWeb-Edu预训练中,DeltaAdamW比AdamW减少46.39%步骤达到相同验证损失(67M模型),增益在1B模型Chinchilla最优预算下持续
  • 理论证明该方法无需矩阵求逆即可实现输入侧曲率校正,且能更快清除陈旧方向

为什么值得看

本文针对深度学习训练中普遍存在的各向异性问题提出了根本性解决方案,将方向感知能力直接嵌入动量更新规则而非外部包装。作为即插即用模块,DeltaMomentum可无缝集成到现有优化器中,在语言模型预训练和视觉任务上均展现出显著效率提升,为优化器设计提供了新的理论视角。

技术解析

  • 核心机制:线性层梯度可分解为输入侧(充当key)和输出侧误差(充当value),DeltaMomentum利用这一Key-Value结构,通过经典Delta规则更新动量缓冲,使每个方向的遗忘速率由其出现频率决定
  • 理论保证:证明该方法构成有效动量,能在固定和漂移最优解下实现输入侧曲率校正而无需矩阵求逆,且清除陈旧方向的速度优于EMA
  • 工程特性:作为任何优化器动量缓冲的即插即用替换,系数在μP下跨宽度可迁移,额外计算开销仅占门控MLP块线性成本的22.2%-25.0%,无持久内存需求
  • 实验验证:在FineWeb-Edu预训练中,DeltaAdamW在67M模型上减少46.39%步骤、370M模型上减少22.12%步骤达到相同验证损失;在Chinchilla最优预算的1B模型上增益持续;Muon基线在相同协议下调优后仍低于DeltaAdamW
  • 泛化能力:增益在SGD、ResNet-18和ViT-Tiny(CIFAR-10)上均成立,训练时诊断数据验证了更好的梯度追踪和健康度更高的输入方向

行业启示

  • 优化器设计范式转变:从"外部包装缓冲"转向"内嵌方向感知",为后续优化器研究提供了新的设计思路,值得在更多架构和任务中验证
  • 预训练效率提升:在语言模型预训练中实现20%-46%的步骤减少,对算力成本敏感的大模型训练具有直接经济价值,建议纳入主流训练流程评估
  • 即插即用架构优势:无需修改优化器核心逻辑即可替换动量缓冲,降低了工程落地门槛,可快速集成到现有训练框架中验证效果

Disclaimer: The above content is generated by AI and is for reference only. 免责声明:以上内容由 AI 生成,仅供参考。

Training 训练 Research 科学研究