ResNets of All Shapes and Sizes: Convergence of Training Dynamics in the Large-scale Limit

• URL: http://arxiv.org/abs/2603.18168v1
• Date: Wed, 18 Mar 2026 18:07:02 GMT
• Title: ResNets of All Shapes and Sizes: Convergence of Training Dynamics in the Large-scale Limit
• Abstract summary: We establish convergence of the training dynamics of residual neural networks (ResNets) to their joint infinite depth L, hidden width M, and embedding dimension D limit.<n>This work completes the program initiated in the companion paper [Chi25] where it is proved that for a fixed embedding dimension D, the training dynamics converges to a Mean ODE model.
• Score: 11.482815450838887
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: We establish convergence of the training dynamics of residual neural networks (ResNets) to their joint infinite depth L, hidden width M, and embedding dimension D limit. Specifically, we consider ResNets with two-layer perceptron blocks in the maximal local feature update (MLU) regime and prove that, after a bounded number of training steps, the error between the ResNet and its large-scale limit is O(1/L + sqrt(D/(L M)) + 1/sqrt(D)). This error rate is empirically tight when measured in embedding space. For a budget of P = Theta(L M D) parameters, this yields a convergence rate O(P^(-1/6)) for the scalings of (L, M, D) that minimize the bound. Our analysis exploits in an essential way the depth-two structure of residual blocks and applies formally to a broad class of state-of-the-art architectures, including Transformers with bounded key-query dimension. From a technical viewpoint, this work completes the program initiated in the companion paper [Chi25] where it is proved that for a fixed embedding dimension D, the training dynamics converges to a Mean ODE dynamics at rate O(1/L + sqrt(D)/sqrt(L M)). Here, we study the large-D limit of this Mean ODE model and establish convergence at rate O(1/sqrt(D)), yielding the above bound by a triangle inequality. To handle the rich probabilistic structure of the limit dynamics and obtain one of the first rigorous quantitative convergence for a DMFT-type limit, we combine the cavity method with propagation of chaos arguments at a functional level on so-called skeleton maps, which express the weight updates as functions of CLT-type sums from the past.

Related papers

• Representational Capacity: Geometric Limits on Feature Representation in Transformer Language Models [51.56484100374058]
  We develop a framework for estimating how many near-orthogonal directions a model can support.<n>Applying this metric across dozens of open-source models reveals two classes: models with high $varepsilon$ whose embeddings lack near-orthogonal structure, and models with low $varepsilon$ that maintain it.
  arXiv  Detail & Related papers  (2026-06-01T18:28:56Z)
• Uniform Scaling Limits in AdamW-Trained Transformers [0.21485350418225238]
  We study the large-depth limit of transformers trained with AdamW.<n>We prove that the joint dynamics of the hidden states and backpropagated variables converge in $L2$.
  arXiv  Detail & Related papers  (2026-05-11T16:54:46Z)
• Provable FDR Control for Deep Feature Selection: Deep MLPs and Beyond [0.0]
  We develop a flexible feature selection framework based on deep neural networks that approximately controls the false discovery rate (FDR), a measure of Type-I error.<n>We show that each coordinate of gradient-based feature vector admits a marginal normal approximation, thereby supporting the validity of FDR control.
  arXiv  Detail & Related papers  (2025-12-04T11:46:06Z)
• The Hidden Width of Deep ResNets: Tight Error Bounds and Phase Diagrams [15.246178589173523]
  We study the gradient-based training of large-depth residual networks (ResNets)<n>We show that with a diverging depth $L$, a fixed embedding dimension $D$, and an arbitrary hidden width $M$, the training dynamics converges to a Neural Mean ODE training dynamics.
  arXiv  Detail & Related papers  (2025-09-12T11:51:44Z)
• Hamiltonian Mechanics of Feature Learning: Bottleneck Structure in Leaky ResNets [58.460298576330835]
  We study Leaky ResNets, which interpolate between ResNets and Fully-Connected nets depending on an 'effective depth'<n>We leverage this intuition to explain the emergence of a bottleneck structure, as observed in previous work.
  arXiv  Detail & Related papers  (2024-05-27T18:15:05Z)
• Learning with Norm Constrained, Over-parameterized, Two-layer Neural Networks [54.177130905659155]
  Recent studies show that a reproducing kernel Hilbert space (RKHS) is not a suitable space to model functions by neural networks. In this paper, we study a suitable function space for over- parameterized two-layer neural networks with bounded norms.
  arXiv  Detail & Related papers  (2024-04-29T15:04:07Z)
• Multi-Grid Tensorized Fourier Neural Operator for High-Resolution PDEs [93.82811501035569]
  We introduce a new data efficient and highly parallelizable operator learning approach with reduced memory requirement and better generalization. MG-TFNO scales to large resolutions by leveraging local and global structures of full-scale, real-world phenomena. We demonstrate superior performance on the turbulent Navier-Stokes equations where we achieve less than half the error with over 150x compression.
  arXiv  Detail & Related papers  (2023-09-29T20:18:52Z)
• A Functional-Space Mean-Field Theory of Partially-Trained Three-Layer Neural Networks [49.870593940818715]
  We study the infinite-width limit of a type of three-layer NN model whose first layer is random and fixed. Our theory accommodates different scaling choices of the model, resulting in two regimes of the MF limit that demonstrate distinctive behaviors.
  arXiv  Detail & Related papers  (2022-10-28T17:26:27Z)
• High-dimensional limit theorems for SGD: Effective dynamics and critical scaling [6.950316788263433]
  We prove limit theorems for the trajectories of summary statistics of gradient descent (SGD) We show a critical scaling regime for the step-size, below which the effective ballistic dynamics matches gradient flow for the population loss. About the fixed points of this effective dynamics, the corresponding diffusive limits can be quite complex and even degenerate.
  arXiv  Detail & Related papers  (2022-06-08T17:42:18Z)
• Global convergence of ResNets: From finite to infinite width using linear parameterization [0.0]
  We study Residual Networks (ResNets) in which the residual block has linear parametrization while still being nonlinear. In this limit, we prove a local Polyak-Lojasiewicz inequality, retrieving the lazy regime. Our analysis leads to a practical and quantified recipe.
  arXiv  Detail & Related papers  (2021-12-10T13:38:08Z)
• Machine Learning For Elliptic PDEs: Fast Rate Generalization Bound, Neural Scaling Law and Minimax Optimality [11.508011337440646]
  We study the statistical limits of deep learning techniques for solving elliptic partial differential equations (PDEs) from random samples. To simplify the problem, we focus on a prototype elliptic PDE: the Schr"odinger equation on a hypercube with zero Dirichlet boundary condition. We establish upper and lower bounds for both methods, which improves upon concurrently developed upper bounds for this problem.
  arXiv  Detail & Related papers  (2021-10-13T17:26:31Z)
• ResNet-LDDMM: Advancing the LDDMM Framework Using Deep Residual Networks [86.37110868126548]
  In this work, we make use of deep residual neural networks to solve the non-stationary ODE (flow equation) based on a Euler's discretization scheme. We illustrate these ideas on diverse registration problems of 3D shapes under complex topology-preserving transformations.
  arXiv  Detail & Related papers  (2021-02-16T04:07:13Z)
• A Dynamical Central Limit Theorem for Shallow Neural Networks [48.66103132697071]
  We prove that the fluctuations around the mean limit remain bounded in mean square throughout training. If the mean-field dynamics converges to a measure that interpolates the training data, we prove that the deviation eventually vanishes in the CLT scaling.
  arXiv  Detail & Related papers  (2020-08-21T18:00:50Z)

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.