Generative foundation models are increasingly scaled in both width and depth, posing significant challenges for stable feature learning and reliable hyperparameter (HP) transfer across model sizes. While maximal update parameterization ($μ$P) has provided a principled solution to both problems for width scaling, existing extensions to the joint width-depth scaling regime remain fragmented, architecture- and optimizer-specific, and often rely on technically involved theories. In this work, we develop a simple and unified spectral framework for $μ$P under joint width-depth scaling. For deep residual networks whose residual blocks contain $k$ transformations, the framework specifies how the norms of weights and their per-step updates should scale with width and depth. It reveals a fundamental transition from $k=1$ to $k\geq 2$, unifying previously disparate $μ$P formulations and identifying the $k\geq 2$ case as more appropriate for practical architectures with multi-transformation branches such as Transformers. Building on this framework, we derive a general recipe for implementing $μ$P across a broad class of optimizers by mapping spectral constraints to concrete HP parameterizations, recovering existing results and extending them to additional optimizers. Finally, experiments on GPT-2 style language models show that the $μ$P formulation derived from the $k\geq 2$ case achieves stable feature learning and robust HP transfer under width-depth scaling, whereas standard parameterization and $μ$P in the $k=1$ case often fail to do so. These results support the practical effectiveness of the proposed spectral framework.
This paper studies the expressive power of graph neural networks falling within the message-passing framework (GNNmp). Two results are presented. First, GNNmp are shown to be Turing universal under sufficient conditions on their depth, width, node attributes, and layer expressiveness. Second, it is discovered that GNNmp can lose a significant portion of their power when their depth and width is restricted. The proposed impossibility statements stem from a new technique that enables the repurposing of seminal results from distributed computing and leads to lower bounds for an array of decision,…
University of Toronto（多伦多大学）；Vector Institute（向量研究所）；University of Guelph（圭尔夫大学）
Theoretical results show that neural networks can be approximated by Gaussian processes in the infinite-width limit. However, for fully connected networks, it has been previously shown that for any fixed network width, $n$, the Gaussian approximation gets worse as the network depth, $d$, increases. Given that modern networks are deep, this raises the question of how well modern architectures, like ResNets, are captured by the infinite-width limit. To provide a better approximation, we study ReLU ResNets in the infinite-depth-and-width limit, where both depth and width tend to infinity as their…
Self-attention architectures, which are rapidly pushing the frontier in natural language processing, demonstrate a surprising depth-inefficient behavior: previous works indicate that increasing the internal representation (network width) is just as useful as increasing the number of self-attention layers (network depth). We theoretically predict a width-dependent transition between depth-efficiency and depth-inefficiency in self-attention. We conduct systematic empirical ablations on networks of depths 6 to 48 that clearly reveal the theoretically predicted behaviors, and provide explicit quan…
In this paper, we study the infinite-depth limit of finite-width residual neural networks with random Gaussian weights. With proper scaling, we show that by fixing the width and taking the depth to infinity, the pre-activations converge in distribution to a zero-drift diffusion process. Unlike the infinite-width limit where the pre-activation converge weakly to a Gaussian random variable, we show that the infinite-depth limit yields different distributions depending on the choice of the activation function. We document two cases where these distributions have closed-form (different) expression…
Texas A&M University（得克萨斯农工大学）；University of Toronto（多伦多大学）
We prove the precise scaling, at finite depth and width, for the mean and variance of the neural tangent kernel (NTK) in a randomly initialized ReLU network. The standard deviation is exponential in the ratio of network depth to width. Thus, even in the limit of infinite overparameterization, the NTK is not deterministic if depth and width simultaneously tend to infinity. Moreover, we prove that for such deep and wide networks, the NTK has a non-trivial evolution during training by showing that the mean of its first SGD update is also exponential in the ratio of network depth to width. This is…
National University of Singapore（新加坡国立大学）；Purdue University（普渡大学）
This paper concentrates on the approximation power of deep feed-forward neural networks in terms of width and depth. It is proved by construction that ReLU networks with width $\mathcal{O}\big(\max\{d\lfloor N^{1/d}\rfloor,\, N+2\}\big)$ and depth $\mathcal{O}(L)$ can approximate a Hölder continuous function on $[0,1]^d$ with an approximation rate $\mathcal{O}\big(λ\sqrt{d} (N^2L^2\ln N)^{-α/d}\big)$, where $α\in (0,1]$ and $λ>0$ are Hölder order and constant, respectively. Such a rate is optimal up to a constant in terms of width and depth separately, while existing results are only nearly op…
While classic studies proved that wide networks allow universal approximation, recent research and successes of deep learning demonstrate the power of deep networks. Based on a symmetric consideration, we investigate if the design of artificial neural networks should have a directional preference, and what the mechanism of interaction is between the width and depth of a network. Inspired by the De Morgan law, we address this fundamental question by establishing a quasi-equivalence between the width and depth of ReLU networks in two aspects. First, we formulate two transforms for mapping an arb…
National University of Singapore（新加坡国立大学）；University of Maryland, College Park（马里兰大学帕克分校）
This paper proposes a new neural network architecture by introducing an additional dimension called height beyond width and depth. Neural network architectures with height, width, and depth as hyper-parameters are called three-dimensional architectures. It is shown that neural networks with three-dimensional architectures are significantly more expressive than the ones with two-dimensional architectures (those with only width and depth as hyper-parameters), e.g., standard fully connected networks. The new network architecture is constructed recursively via a nested structure, and hence we call…
Stanford University（斯坦福大学）；Singapore University of Technology and Design（新加坡科技设计大学）
The expressivity of neural networks as a function of their depth, width and type of activation units has been an important question in deep learning theory. Recently, depth separation results for ReLU networks were obtained via a new connection with dynamical systems, using a generalized notion of fixed points of a continuous map $f$, called periodic points. In this work, we strengthen the connection with dynamical systems and we improve the existing width lower bounds along several aspects. Our first main result is period-specific width lower bounds that hold under the stronger notion of $L^1…
We provide several new depth-based separation results for feed-forward neural networks, proving that various types of simple and natural functions can be better approximated using deeper networks than shallower ones, even if the shallower networks are much larger. This includes indicators of balls and ellipses; non-linear functions which are radial with respect to the $L_1$ norm; and smooth non-linear functions. We also show that these gaps can be observed experimentally: Increasing the depth indeed allows better learning than increasing width, when training neural networks to learn an indicat…
We solve an open question from Lu et al. (2017), by showing that any target network with inputs in $\mathbb{R}^d$ can be approximated by a width $O(d)$ network (independent of the target network's architecture), whose number of parameters is essentially larger only by a linear factor. In light of previous depth separation theorems, which imply that a similar result cannot hold when the roles of width and depth are interchanged, it follows that depth plays a more significant role than width in the expressive power of neural networks. We extend our results to constructing networks with bounded w…
Neural Networks and Deep Learning lab.（神经网络与深度学习实验室）；Moscow Institute of Physics and Technology（莫斯科物理技术学院）
Recent research has been focused on two different approaches to studying neural networks training in the limit of infinite width (1) a mean-field (MF) and (2) a constant neural tangent kernel (NTK) approximations. These two approaches have different scaling of hyperparameters with the width of a network layer and as a result, different infinite-width limit models. We propose a general framework to study how the limit behavior of neural models depends on the scaling of hyperparameters with network width. Our framework allows us to derive scaling for existing MF and NTK limits, as well as an unc…
We derive finite width and depth corrections for the Neural Tangent Kernel (NTK) of ResNets and DenseNets. Our analysis reveals that finite size residual architectures are initialized much closer to the "kernel regime" than their vanilla counterparts: while in networks that do not use skip connections, convergence to the NTK requires one to fix the depth, while increasing the layers' width. Our findings show that in ResNets, convergence to the NTK may occur when depth and width simultaneously tend to infinity, provided with a proper initialization. In DenseNets, however, convergence of the NTK…
A key factor in the success of deep neural networks is the ability to scale models to improve performance by varying the architecture depth and width. This simple property of neural network design has resulted in highly effective architectures for a variety of tasks. Nevertheless, there is limited understanding of effects of depth and width on the learned representations. In this paper, we study this fundamental question. We begin by investigating how varying depth and width affects model hidden representations, finding a characteristic block structure in the hidden representations of larger c…
Deep Linear Network Training Dynamics from Random Initialization: Data, Width, Depth, and Hyperparameter Transfer
John Paulson School of Engineering and Applied Sciences, Harvard University（哈佛大学约翰·保罗森工程与应用科学学校）；Center for Brain Sciences（脑科学研究中心）；Kempner Institute（凯普纳研究所）
Singapore University of Technology and Design（新加坡科技设计大学）；Stanford University（斯坦福大学）；SUTD（新加坡科技设计大学）
Understanding the representational power of Deep Neural Networks (DNNs) and how their structural properties (e.g., depth, width, type of activation unit) affect the functions they can compute, has been an important yet challenging question in deep learning and approximation theory. In a seminal paper, Telgarsky highlighted the benefits of depth by presenting a family of functions (based on simple triangular waves) for which DNNs achieve zero classification error, whereas shallow networks with fewer than exponentially many nodes incur constant error. Even though Telgarsky's work reveals the lim…
Theory of Scaling Laws for In-Context Regression: Depth, Width, Context and Time
Center for Mathematical Sciences and Applications（数学科学与应用中心）；John A. Paulson School of Engineering and Applied Sciences（约翰·A·保罗森工程与应用科学学院）；Kempner Institute for the Study of Natural and Artificial Intelligence（自然与人工智能研究学院）；Harvard University（哈佛大学）
Neural Tangent Kernel (NTK) is widely used to analyze overparametrized neural networks due to the famous result by Jacot et al. (2018): in the infinite-width limit, the NTK is deterministic and constant during training. However, this result cannot explain the behavior of deep networks, since it generally does not hold if depth and width tend to infinity simultaneously. In this paper, we study the NTK of fully-connected ReLU networks with depth comparable to width. We prove that the NTK properties depend significantly on the depth-to-width ratio and the distribution of parameters at initializat…
In this paper, we analyze the effects of depth and width on the quality of local minima, without strong over-parameterization and simplification assumptions in the literature. Without any simplification assumption, for deep nonlinear neural networks with the squared loss, we theoretically show that the quality of local minima tends to improve towards the global minimum value as depth and width increase. Furthermore, with a locally-induced structure on deep nonlinear neural networks, the values of local minima of neural networks are theoretically proven to be no worse than the globally optimal …
One of the central problems in the study of deep learning theory is to understand how the structure properties, such as depth, width and the number of nodes, affect the expressivity of deep neural networks. In this work, we show a new connection between the expressivity of deep neural networks and topological entropy from dynamical system, which can be used to characterize depth-width trade-offs of neural networks. We provide an upper bound on the topological entropy of neural networks with continuous semi-algebraic units by the structure parameters. Specifically, the topological entropy of Re…
We study the average robustness notion in deep neural networks in (selected) wide and narrow, deep and shallow, as well as lazy and non-lazy training settings. We prove that in the under-parameterized setting, width has a negative effect while it improves robustness in the over-parameterized setting. The effect of depth closely depends on the initialization and the training mode. In particular, when initialized with LeCun initialization, depth helps robustness with the lazy training regime. In contrast, when initialized with Neural Tangent Kernel (NTK) and He-initialization, depth hurts the ro…
A Closed Form Depth-Width Law for Neural Collapse in Deep Linear Networks
Institute of AI for Health, Helmholtz Munich（亥姆霍兹慕尼黑人工智能健康研究所）；Université de Montréal（蒙特利尔大学）；Mila – Quebec AI Institute（米拉-魁北克人工智能研究所）；Faculty of Mathematics, Informatics and Mechanics, University of Warsaw（华沙大学数学、信息与力学学院）；Faculty of Mathematics, Informatics and Statistics, LMU Munich（慕尼黑大学数学、信息与统计学院）；School of Computation, Information and Technology, Technical University of Munich（慕尼黑工业大学计算、信息与技术学院）
Peking University（北京大学）；Key Laboratory of Machine Perception, MOE, School of EECS, Peking University（北京大学信息科学技术学院机器感知与智能教育部重点实验室）；Center for Data Science, Peking University（北京大学数据科学研究中心）；Beijing Institute of Big Data Research（北京大数据研究院）
The expressive power of neural networks is important for understanding deep learning. Most existing works consider this problem from the view of the depth of a network. In this paper, we study how width affects the expressiveness of neural networks. Classical results state that depth-bounded (e.g. depth-$2$) networks with suitable activation functions are universal approximators. We show a universal approximation theorem for width-bounded ReLU networks: width-$(n+4)$ ReLU networks, where $n$ is the input dimension, are universal approximators. Moreover, except for a measure zero set, all funct…
The pre-trained language models like BERT, though powerful in many natural language processing tasks, are both computation and memory expensive. To alleviate this problem, one approach is to compress them for specific tasks before deployment. However, recent works on BERT compression usually compress the large BERT model to a fixed smaller size. They can not fully satisfy the requirements of different edge devices with various hardware performances. In this paper, we propose a novel dynamic BERT model (abbreviated as DynaBERT), which can flexibly adjust the size and latency by selecting adapti…
Mixture of Universal Experts: Scaling Virtual Width via Depth-Width Transformation
Institute of Information Engineering, Chinese Academy of Sciences（中国科学院信息工程研究所）；School of Cyber Security, University of Chinese Academy of Sciences（中国科学院大学网络安全学院）；Baidu Inc（百度公司）
Cornell University（康奈尔大学）；University of Connecticut（康涅狄格大学）；University of California, Los Angeles（加利福尼亚大学洛杉矶分校）
The standard Universal Approximation Theorem for operator neural networks (NNs) holds for arbitrary width and bounded depth. Here, we prove that operator NNs of bounded width and arbitrary depth are universal approximators for continuous nonlinear operators. In our main result, we prove that for non-polynomial activation functions that are continuously differentiable at a point with a nonzero derivative, one can construct an operator NN of width five, whose inputs are real numbers with finite decimal representations, that is arbitrarily close to any given continuous nonlinear operator. We deri…
SkyView Academy（SkyView学院）；University of Colorado Boulder（科罗拉多大学博尔德分校）
Neural network width and depth are fundamental aspects of network topology. Universal approximation theorems provide that with increasing width or depth, there exists a neural network that approximates a function arbitrarily well. These theorems assume requirements, such as infinite data, that must be discretized in practice. Through numerical experiments, we seek to test the lower bounds established by Hanin in 2017.
Department of Mathematics, Texas A\&M, College Station, United States
This article concerns the expressive power of depth in neural nets with ReLU activations and bounded width. We are particularly interested in the following questions: what is the minimal width $w_{\text{min}}(d)$ so that ReLU nets of width $w_{\text{min}}(d)$ (and arbitrary depth) can approximate any continuous function on the unit cube $[0,1]^d$ aribitrarily well? For ReLU nets near this minimal width, what can one say about the depth necessary to approximate a given function? Our approach to this paper is based on the observation that, due to the convexity of the ReLU activation, ReLU nets a…
Depth-Width tradeoffs in Algorithmic Reasoning of Graph Tasks with Transformers
Courant Institute of Mathematical Sciences, New York University（纽约大学应用数学科学研究所）；Google Research（谷歌研究）；Meta AI；Bar-Ilan University（巴伊兰大学）；Department of Bio-Medical Engineering, Edmond J. Safra Center for Bioinformatics, Tel-Aviv University（生物医学工程系，埃德蒙·J·萨法中心，特拉维夫大学）；Tel Aviv University（特拉维夫大学）
In this note, we first derive a one-parameter family of hyperparameter scaling strategies that interpolates between the neural-tangent scaling and mean-field/maximal-update scaling. We then calculate the scalings of dynamical observables -- network outputs, neural tangent kernels, and differentials of neural tangent kernels -- for wide and deep neural networks. These calculations in turn reveal a proper way to scale depth with width such that resultant large-scale models maintain their representation-learning ability. Finally, we observe that various infinite-width limits examined in the liter…
Implicit deep learning has recently become popular in the machine learning community since these implicit models can achieve competitive performance with state-of-the-art deep networks while using significantly less memory and computational resources. However, our theoretical understanding of when and how first-order methods such as gradient descent (GD) converge on \textit{nonlinear} implicit networks is limited. Although this type of problem has been studied in standard feed-forward networks, the case of implicit models is still intriguing because implicit networks have \textit{infinitely} m…
Deep neural networks have been used in various machine learning applications and achieved tremendous empirical successes. However, training deep neural networks is a challenging task. Many alternatives have been proposed in place of end-to-end back-propagation. Layer-wise training is one of them, which trains a single layer at a time, rather than trains the whole layers simultaneously. In this paper, we study a layer-wise training using a block coordinate gradient descent (BCGD) for deep linear networks. We establish a general convergence analysis of BCGD and found the optimal learning rate, w…
Department of Mathematics, Texas A&M（德克萨斯农工大学数学系）；Trinity College, Cambridge（剑桥大学三一学院）
This article concerns the expressive power of depth in deep feed-forward neural nets with ReLU activations. Specifically, we answer the following question: for a fixed $d_{in}\geq 1,$ what is the minimal width $w$ so that neural nets with ReLU activations, input dimension $d_{in}$, hidden layer widths at most $w,$ and arbitrary depth can approximate any continuous, real-valued function of $d_{in}$ variables arbitrarily well? It turns out that this minimal width is exactly equal to $d_{in}+1.$ That is, if all the hidden layer widths are bounded by $d_{in}$, then even in the infinite depth limit…
Beyond Scaling Curves: Internal Dynamics of Neural Networks Through the NTK Lens
Institute of Computational Physics, University of Stuttgart, Stuttgart, Germany（计算物理研究所，斯图加特大学，斯图加特，德国）；Cavendish Laboratory（卡文迪许实验室）；DAMTP, University of Cambridge, Cambridge, United Kingdom, CB3 0WA（戴维森数学理论物理系，剑桥大学，剑桥，英国，CB3 0WA）
Memorization capacity of deep ReLU neural networks characterized by width and depth
School of Mathematics (Zhuhai), Sun Yat-sen University, Zhuhai, China（中山大学数学学院（珠海）、中山大学珠海校区，中国）；School of Mathematics (Zhuhai) and Guangdong Province Key Laboratory of Computational Science, Sun Yat-sen University, Zhuhai, China（中山大学数学学院（珠海）和广东省计算科学重点实验室，中山大学珠海校区，中国）
Robust Layerwise Scaling Rules by Proper Weight Decay Tuning
Department of Electrical Engineering and Computer Science, MIT, Cambridge, MA, USA（麻省理工学院电子工程与计算机科学系）；Department of Computer Science, UCLA, CA, USA（加州大学洛杉矶分校计算机科学系）
Optimal Learning Rate Scaling Depends on Data in Deep Scalar Linear Networks
Gatsby Computational Neuroscience Unit, University College London（加布兹计算神经科学单位，伦敦大学学院）；Sainsbury Wellcome Centre, University College London（萨金斯·惠康中心，伦敦大学学院）
National University of Singapore（新加坡国立大学）；Purdue University（普渡大学）
This paper establishes the (nearly) optimal approximation error characterization of deep rectified linear unit (ReLU) networks for smooth functions in terms of both width and depth simultaneously. To that end, we first prove that multivariate polynomials can be approximated by deep ReLU networks of width $\mathcal{O}(N)$ and depth $\mathcal{O}(L)$ with an approximation error $\mathcal{O}(N^{-L})$. Through local Taylor expansions and their deep ReLU network approximations, we show that deep ReLU networks of width $\mathcal{O}(N\ln N)$ and depth $\mathcal{O}(L\ln L)$ can approximate $f\in C^s([0…
National University of Singapore（新加坡国立大学）；Purdue University（普渡大学）
A new network with super approximation power is introduced. This network is built with Floor ($\lfloor x\rfloor$) or ReLU ($\max\{0,x\}$) activation function in each neuron and hence we call such networks Floor-ReLU networks. For any hyper-parameters $N\in\mathbb{N}^+$ and $L\in\mathbb{N}^+$, it is shown that Floor-ReLU networks with width $\max\{d,\, 5N+13\}$ and depth $64dL+3$ can uniformly approximate a Hölder function $f$ on $[0,1]^d$ with an approximation error $3λd^{α/2}N^{-α\sqrt{L}}$, where $α\in(0,1]$ and $λ$ are the Hölder order and constant, respectively. More generally for an arbit…
