Dynamics of Stochastic Momentum with Sparse Updates in High Dimensions

• URL: http://arxiv.org/abs/2605.28961v1
• Date: Wed, 27 May 2026 18:06:41 GMT
• Title: Dynamics of Stochastic Momentum with Sparse Updates in High Dimensions
• Abstract summary: Existing theory of momentum assumes that gradients arrive at every parameter at a roughly constant rate.<n>We theoretically analyze the dynamics of two tractable models of momentum under sparse updates.
• Score: 5.674162165646654
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Existing theory of momentum assumes that gradients arrive at every parameter at a roughly constant rate, an assumption violated in practice by heavy-tailed data distributions and modern architectures. We theoretically analyze the dynamics of two tractable models of momentum under sparse updates: a least squares model with sparse inputs and a logistic regression model with a rare class. Both admit exact closed-form second-moment dynamics whose high-dimensional limits we characterize across three scaling exponents for sparsity, batch size, and momentum decay. The phase structure on both problems is governed by the ratio of two intrinsic timescales: a momentum retention timescale (how many active updates the buffer survives) and a learning timescale (how many active updates it takes to reduce the squared error). When learning is much slower than retention, the limit matches SGD; when learning is faster, the system is unstable; where the timescales coincide, we recover classical heavy-ball dynamics. The oscillatory dynamics occur at different momentum values for different token sparsity, creating a spectral conflict for global momentum across token frequencies.

Related papers

• Transformer Learning of Chaotic Collective Dynamics in Many-Body Systems [0.0]
  We show that a self-attention-based transformer framework provides an effective approach for modeling chaotic collective dynamics.<n>We study the one-dimensional semiclassical Holstein model, where interaction quenches induce strongly nonlinear and chaotic dynamics.<n>Our results establish self-attention as a powerful mechanism for learning effective reduced dynamics in chaotic many-body systems.
  arXiv  Detail & Related papers  (2026-01-27T01:33:33Z)
• Sparse-to-Field Reconstruction via Stochastic Neural Dynamic Mode Decomposition [12.812771670043212]
  Many real-world systems, like wind fields and ocean currents, are dynamic and hard to model.<n> Dynamic Mode Decomposition (DMD) provides a simple, data-driven approximation, but practical use is limited by sparse/noisy observations.<n>We introduce NODE-DMD, a probabilistic extension of DMD that models continuous-time, nonlinear dynamics while remaining interpretable.
  arXiv  Detail & Related papers  (2025-11-25T18:39:50Z)
• Forecasting Continuous Non-Conservative Dynamical Systems in SO(3) [51.510040541600176]
  We propose a novel approach to modeling the rotation of moving objects in computer vision.<n>Our approach is agnostic to energy and momentum conservation while being robust to input noise.<n>By learning to approximate object dynamics from noisy states during training, our model attains robust extrapolation capabilities in simulation and various real-world settings.
  arXiv  Detail & Related papers  (2025-08-11T09:03:10Z)
• Langevin Flows for Modeling Neural Latent Dynamics [81.81271685018284]
  We introduce LangevinFlow, a sequential Variational Auto-Encoder where the time evolution of latent variables is governed by the underdamped Langevin equation.<n>Our approach incorporates physical priors -- such as inertia, damping, a learned potential function, and forces -- to represent both autonomous and non-autonomous processes in neural systems.<n>Our method outperforms state-of-the-art baselines on synthetic neural populations generated by a Lorenz attractor.
  arXiv  Detail & Related papers  (2025-07-15T17:57:48Z)
• Scaling Collapse Reveals Universal Dynamics in Compute-Optimally Trained Neural Networks [59.552873049024775]
  We show that compute-optimally trained models exhibit a remarkably precise universality.<n>With learning rate decay, the collapse becomes so tight that differences in the normalized curves across models fall below the noise floor.<n>We explain these phenomena by connecting collapse to the power-law structure in typical neural scaling laws.
  arXiv  Detail & Related papers  (2025-07-02T20:03:34Z)
• Intensity Profile Projection: A Framework for Continuous-Time Representation Learning for Dynamic Networks [50.2033914945157]
  We present a representation learning framework, Intensity Profile Projection, for continuous-time dynamic network data. The framework consists of three stages: estimating pairwise intensity functions, learning a projection which minimises a notion of intensity reconstruction error. Moreoever, we develop estimation theory providing tight control on the error of any estimated trajectory, indicating that the representations could even be used in quite noise-sensitive follow-on analyses.
  arXiv  Detail & Related papers  (2023-06-09T15:38:25Z)
• Flatter, faster: scaling momentum for optimal speedup of SGD [0.0]
  We study training dynamics arising from interplay between gradient descent (SGD) and label noise and momentum in the training of neural networks. We find that scaling the momentum hyper parameter $1-NISTbeta$ with the learning rate to the power of $2/3$ maximally accelerates training, without sacrificing generalization.
  arXiv  Detail & Related papers  (2022-10-28T20:41:48Z)
• Learning Physical Dynamics with Subequivariant Graph Neural Networks [99.41677381754678]
  Graph Neural Networks (GNNs) have become a prevailing tool for learning physical dynamics. Physical laws abide by symmetry, which is a vital inductive bias accounting for model generalization. Our model achieves on average over 3% enhancement in contact prediction accuracy across 8 scenarios on Physion and 2X lower rollout MSE on RigidFall.
  arXiv  Detail & Related papers  (2022-10-13T10:00:30Z)
• Losing momentum in continuous-time stochastic optimisation [42.617042045455506]
  momentum-based optimisation algorithms have become particularly widespread. In this work, we analyse a continuous-time model for gradient descent with momentum. We also train a convolutional neural network in an image classification problem.
  arXiv  Detail & Related papers  (2022-09-08T10:46:05Z)
• Dynamics with autoregressive neural quantum states: application to critical quench dynamics [41.94295877935867]
  We present an alternative general scheme that enables one to capture long-time dynamics of quantum systems in a stable fashion. We apply the scheme to time-dependent quench dynamics by investigating the Kibble-Zurek mechanism in the two-dimensional quantum Ising model.
  arXiv  Detail & Related papers  (2022-09-07T15:50:00Z)
• Multivariate Time Series Forecasting with Dynamic Graph Neural ODEs [65.18780403244178]
  We propose a continuous model to forecast Multivariate Time series with dynamic Graph neural Ordinary Differential Equations (MTGODE) Specifically, we first abstract multivariate time series into dynamic graphs with time-evolving node features and unknown graph structures. Then, we design and solve a neural ODE to complement missing graph topologies and unify both spatial and temporal message passing.
  arXiv  Detail & Related papers  (2022-02-17T02:17:31Z)

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.