Every Layer Counts: An Exponential $L_2$ Depth Hierarchy for ReLU Networks

• URL: http://arxiv.org/abs/2608.23877v2
• Date: Sun, 30 Aug 2026 12:24:04 GMT
• Title: Every Layer Counts: An Exponential $L_2$ Depth Hierarchy for ReLU Networks
• Abstract summary: We prove a depth hierarchy for ReLU neural networks in which every additional ReLU layer can save exponentially many neurons.<n>We also prove an exact separation for a more regular target, which is globally $[0,1]$-valued and $mathcalO(sqrt d)$-cubeschitz.
• Score: 3.0120086446979877
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: We prove a depth hierarchy for ReLU neural networks in which every additional ReLU layer can save exponentially many neurons. For all $k\geq2$, we construct a globally $[0,1]$-valued, $1$-Lipschitz function realized by a depth-$(k+1)$ network of width $\mathcal{O}(d^4)$, whereas any depth-$k$ network with unrestricted weights and width at most $\frac{2^d}{2d(k-1)}$ has squared $L_2$ error at least $1/24$ under an absolutely continuous distribution supported at exponential distance from the origin. To the best of our knowledge, this is the first exponential hierarchy across all adjacent fixed depths, and the first exponential separation for ReLU networks between two fixed depths whose shallower network has depth at least $3$. The lower bound also immediately yields the corresponding hierarchy for exact computation. Moreover, the case $k=2$ gives a compactly supported separation between depths $3$ and $2$ with unrestricted shallow-network weights, answering a question raised by Safran, Eldan, and Shamir (2019). The distribution used in our construction nevertheless has all its mass at exponential radius, placing the hierarchy outside the regularity regime in which such a separation would imply major threshold-circuit lower bounds. We also prove an exact separation for a more regular target, which is globally $[0,1]$-valued and $\mathcal{O}(\sqrt d)$-Lipschitz and maps the unit hypercube onto $[0,1]$. It is computed by a polynomial-width depth-$4$ network, whereas any depth-$3$ network agreeing with it on the unit hypercube requires exponentially many first-layer neurons, even with unrestricted weights.

Related papers

• The Value of Depth in Message Passing on Sparse Graphs: A Kesten-Stigum Dichotomy [0.0]
  We study how deep does a graph neural network need to be on a sparse graph?<n>We prove that the value of depth is governed by a single number, the Kesten-Stigum ratio $=2$.<n>No local classifier of any depth beats the universal floor $e-(-)$ set by isolated roots.
  arXiv  Detail & Related papers  (2026-07-18T07:11:45Z)
• A law of robustness for two-layer neural networks with arbitrary weights [0.0]
  Bubeck, Li and Nagaraj conjectured that, for generic data, any two-layer neural network with $m$ neurons that fits noisy labels must have Lipschitz constant.<n>We prove the conjectured law, up to one logarithmic factor, for every continuous piecewise-linear activation in particular for ReLU networks.
  arXiv  Detail & Related papers  (2026-07-08T17:44:15Z)
• Geometric Layer-wise Approximation Rates for Deep Networks [13.496991650323038]
  We develop a framework in which depth admits a precise scale-dependent interpretation.<n>Our network design is inspired by multigrade deep learning, where depth serves as a progressive refinement mechanism.
  arXiv  Detail & Related papers  (2026-04-22T06:09:20Z)
• Constructive Universal Approximation and Finite Sample Memorization by Narrow Deep ReLU Networks [0.0]
  We show that any dataset with $N$ distinct points in $mathbbRd$ and $M$ output classes can be exactly classified.<n>We also prove a universal approximation theorem in $Lp(Omega; mathbbRm)$ for any bounded domain.<n>Our results offer a unified and interpretable framework connecting controllability, expressivity, and training dynamics in deep neural networks.
  arXiv  Detail & Related papers  (2024-09-10T14:31:21Z)
• Implicit Hypersurface Approximation Capacity in Deep ReLU Networks [0.0]
  We develop a geometric approximation theory for deep feed-forward neural networks with ReLU activations. We show that a deep fully-connected ReLU network of width $d+1$ can implicitly construct an approximation as its zero contour.
  arXiv  Detail & Related papers  (2024-07-04T11:34:42Z)
• Bayesian Inference with Deep Weakly Nonlinear Networks [57.95116787699412]
  We show at a physics level of rigor that Bayesian inference with a fully connected neural network is solvable. We provide techniques to compute the model evidence and posterior to arbitrary order in $1/N$ and at arbitrary temperature.
  arXiv  Detail & Related papers  (2024-05-26T17:08:04Z)
• Learning Hierarchical Polynomials with Three-Layer Neural Networks [56.71223169861528]
  We study the problem of learning hierarchical functions over the standard Gaussian distribution with three-layer neural networks. For a large subclass of degree $k$s $p$, a three-layer neural network trained via layerwise gradientp descent on the square loss learns the target $h$ up to vanishing test error. This work demonstrates the ability of three-layer neural networks to learn complex features and as a result, learn a broad class of hierarchical functions.
  arXiv  Detail & Related papers  (2023-11-23T02:19:32Z)
• On Expressivity of Height in Neural Networks [29.49793694185358]
  We call a neural network characterized by width, depth, and height a 3D network.<n>We show via bound estimation and explicit construction that given the same number of neurons and parameters, a 3D ReLU network of width $W$, depth $K$, and height $H$ has greater expressive power than a 2D network of width $Htimes W$ and depth $K$.
  arXiv  Detail & Related papers  (2023-05-11T11:54:36Z)
• Understanding Deep Neural Function Approximation in Reinforcement Learning via $\epsilon$-Greedy Exploration [53.90873926758026]
  This paper provides a theoretical study of deep neural function approximation in reinforcement learning (RL) We focus on the value based algorithm with the $epsilon$-greedy exploration via deep (and two-layer) neural networks endowed by Besov (and Barron) function spaces. Our analysis reformulates the temporal difference error in an $L2(mathrmdmu)$-integrable space over a certain averaged measure $mu$, and transforms it to a generalization problem under the non-iid setting.
  arXiv  Detail & Related papers  (2022-09-15T15:42:47Z)
• Shallow neural network representation of polynomials [91.3755431537592]
  We show that $d$-variables of degreeR$ can be represented on $[0,1]d$ as shallow neural networks of width $d+1+sum_r=2Rbinomr+d-1d-1d-1[binomr+d-1d-1d-1[binomr+d-1d-1d-1[binomr+d-1d-1d-1d-1[binomr+d-1d-1d-1d-1
  arXiv  Detail & Related papers  (2022-08-17T08:14:52Z)
• Expressive power of binary and ternary neural networks [91.3755431537592]
  We show that deep sparse ReLU networks with ternary weights and deep ReLU networks with binary weights can approximate $beta$-H"older functions on $[0,1]d$.
  arXiv  Detail & Related papers  (2022-06-27T13:16:08Z)
• On minimal representations of shallow ReLU networks [0.0]
  We show that the minimal representation for $f$ uses either $n$, $n+1$ or $n+2$ neurons. In particular, where the input layer is one-dimensional, minimal representations always use at most $n+1$ neurons but in all higher dimensional settings there are functions for which $n+2$ neurons are needed.
  arXiv  Detail & Related papers  (2021-08-12T10:22:24Z)
• Learning Over-Parametrized Two-Layer ReLU Neural Networks beyond NTK [58.5766737343951]
  We consider the dynamic of descent for learning a two-layer neural network. We show that an over-parametrized two-layer neural network can provably learn with gradient loss at most ground with Tangent samples.
  arXiv  Detail & Related papers  (2020-07-09T07:09:28Z)
• Sharp Representation Theorems for ReLU Networks with Precise Dependence on Depth [26.87238691716307]
  We prove sharp-free representation results for neural networks with $D$ ReLU layers under square loss. Our results confirm the prevailing hypothesis that deeper networks are better at representing less smooth functions.
  arXiv  Detail & Related papers  (2020-06-07T05:25:06Z)

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.