The choice of activation function fundamentally shapes the representational capacity and parameter efficiency of deep neural networks, yet most widely used activations lack rigorous theoretical guarantees on these properties. We provide a theoretical analysis of DeepBern-Nets (DBNs) -- networks employing learnable Bernstein polynomial activations -- showing that their approximation error decays with the network depth $L$ and the polynomial order $n$ with a rate of $\mathcal{O}(n^{-L})$, exponentially faster than the polynomial rate of ReLU architectures while remaining fully differentiable. We validate these predictions through $1{,}344$ experiments on large scientific datasets (HIGGS and SUSY), comparing DBNs against ReLU, Leaky ReLU, SELU, and GeLU. DBNs achieve over $70\%$ parameter reduction across the majority of architectures -- reaching $99.9\%$ at scale -- converge to ReLU's final loss in as few as $26\%$ of the training epochs, and attain up to $45\%$ lower final loss. These advantages hold over all tested activations, confirming that DBN's gains stem from the learnable polynomial structure rather than mere smoothness.
We introduce SPLASH units, a class of learnable activation functions shown to simultaneously improve the accuracy of deep neural networks while also improving their robustness to adversarial attacks. SPLASH units have both a simple parameterization and maintain the ability to approximate a wide range of non-linear functions. SPLASH units are: 1) continuous; 2) grounded (f(0) = 0); 3) use symmetric hinges; and 4) the locations of the hinges are derived directly from the data (i.e. no learning required). Compared to nine other learned and fixed activation functions, including ReLU and its varian…
University of Waterloo（滑铁卢大学）；Waterloo Artificial Intelligence Institute（滑铁卢人工智能研究所）；DarwinAI Corp.（达尔文人工智能公司）
From fully connected neural networks to convolutional neural networks, the learned parameters within a neural network have been primarily relegated to the linear parameters (e.g., convolutional filters). The non-linear functions (e.g., activation functions) have largely remained, with few exceptions in recent years, parameter-less, static throughout training, and seen limited variation in design. Largely ignored by the deep learning community, radial basis function (RBF) networks provide an interesting mechanism for learning more complex non-linear activation functions in addition to the linea…
Qualcomm AI Research, Qualcomm Technologies, Inc.（高通AI研究，高通技术公司）；Qualcomm AI Research, Qualcomm Technologies Netherlands B.V.（高通AI研究，高通技术荷兰有限公司）；Seoul National University（首尔大学）
Unlike ReLU, newer activation functions (like Swish, H-swish, Mish) that are frequently employed in popular efficient architectures can also result in negative activation values, with skewed positive and negative ranges. Typical learnable quantization schemes [PACT, LSQ] assume unsigned quantization for activations and quantize all negative activations to zero which leads to significant loss in performance. Naively using signed quantization to accommodate these negative values requires an extra sign bit which is expensive for low-bit (2-, 3-, 4-bit) quantization. To solve this problem, we prop…
Fuzhou University（福州大学）；University of Vienna（维也纳大学）
In practical applications, multi-view data depicting objectives from assorted perspectives can facilitate the accuracy increase of learning algorithms. However, given multi-view data, there is limited work for learning discriminative node relationships and graph information simultaneously via graph convolutional network that has drawn the attention from considerable researchers in recent years. Most of existing methods only consider the weighted sum of adjacency matrices, yet a joint neural network of both feature and graph fusion is still under-explored. To cope with these issues, this paper …
Applying neural networks as controllers in dynamical systems has shown great promises. However, it is critical yet challenging to verify the safety of such control systems with neural-network controllers in the loop. Previous methods for verifying neural network controlled systems are limited to a few specific activation functions. In this work, we propose a new reachability analysis approach based on Bernstein polynomials that can verify neural-network controlled systems with a more general form of activation functions, i.e., as long as they ensure that the neural networks are Lipschitz conti…
Deep neural networks are known to be vulnerable to adversarially perturbed inputs. A commonly used defense is adversarial training, whose performance is influenced by model capacity. While previous works have studied the impact of varying model width and depth on robustness, the impact of increasing capacity by using learnable parametric activation functions (PAFs) has not been studied. We study how using learnable PAFs can improve robustness in conjunction with adversarial training. We first ask the question: how should we incorporate parameters into activation functions to improve robustness…
Aalen University（阿伦大学）；Óbuda University（奥布达大学）；SupplyPike（SupplyPike公司）；University of Texas at El Paso（得克萨斯大学埃尔帕索分校）；University of Stuttgart（斯图加特大学）
We present a deep learning model for finding human-understandable connections between input features. Our approach uses a parameterized, differentiable activation function, based on the theoretical background of nilpotent fuzzy logic and multi-criteria decision-making (MCDM). The learnable parameter has a semantic meaning indicating the level of compensation between input features. The neural network determines the parameters using gradient descent to find human-understandable relationships between input features. We demonstrate the utility and effectiveness of the model by successfully applyi…
Purdue University（普渡大学）；Michigan State University（密歇根州立大学）；Texas Tech University（德克萨斯理工大学）
We propose reproducing activation functions (RAFs) to improve deep learning accuracy for various applications ranging from computer vision to scientific computing. The idea is to employ several basic functions and their learnable linear combination to construct neuron-wise data-driven activation functions for each neuron. Armed with RAFs, neural networks (NNs) can reproduce traditional approximation tools and, therefore, approximate target functions with a smaller number of parameters than traditional NNs. In NN training, RAFs can generate neural tangent kernels (NTKs) with a better condition …
Element-wise activation functions play a critical role in deep neural networks via affecting the expressivity power and the learning dynamics. Learning-based activation functions have recently gained increasing attention and success. We propose a new perspective of learnable activation function through formulating them with element-wise attention mechanism. In each network layer, we devise an attention module which learns an element-wise, sign-based attention map for the pre-activation feature map. The attention map scales an element based on its sign. Adding the attention module with a rectif…
In this paper, we point out that differential architecture search (DARTS) makes gradient of architecture parameters biased for network weights and architecture parameters are updated in different datasets alternatively in the bi-level optimization framework. The bias causes the architecture parameters of non-learnable operations to surpass that of learnable operations. Moreover, using softmax as architecture parameters' activation function and inappropriate learning rate would exacerbate the bias. As a result, it's frequently observed that non-learnable operations are dominated in the search p…
Université Savoie Mont Blanc（萨瓦大学）；TLECOM Bretagne（布列塔尼电信学院）
The paper proposes representation functionals in a dual paradigm where learning jointly concerns both linear convolutional weights and parametric forms of nonlinear activation functions. The nonlinear forms proposed for performing the functional representation are associated with a new class of parametric neural transfer functions called rectified power sigmoid units. This class is constructed to integrate both advantages of sigmoid and rectified linear unit functions, in addition with rejecting the drawbacks of these functions. Moreover, the analytic form of this new neural class involves sca…
In neural networks literature, there is a strong interest in identifying and defining activation functions which can improve neural network performance. In recent years there has been a renovated interest of the scientific community in investigating activation functions which can be trained during the learning process, usually referred to as "trainable", "learnable" or "adaptable" activation functions. They appear to lead to better network performance. Diverse and heterogeneous models of trainable activation function have been proposed in the literature. In this paper, we present a survey of t…
Department of Physics and Center for Theoretical Physics, National Taiwan University（物理系与理论物理中心，国立台湾大学）；Center for Quantum Science and Engineering, National Taiwan University（量子科学与工程中心，国立台湾大学）；Physics Division, National Center for Theoretical Sciences, National Taiwan University（理论科学国家中心物理组，国立台湾大学）；Department of Computer Science, University of North Carolina at Chapel Hill（计算机科学系，北卡罗来纳大学教堂山分校）
Müntz-Szász Networks: Neural Architectures with Learnable Power-Law Bases
Axiom Research Group（Axiom研究组）；Department of Applied Mathematics and Computational Science, The Nelson Mandela African Institution of Science and Technology (NM-AIST)（应用数学与计算科学系，纳尔逊·曼德拉非洲科学技术研究所）；African Institute for Mathematical Sciences (AIMS), Research and Innovation Centre (RIC)（非洲数学科学研究所，研究与创新中心）
High-Frequency (HF) signals are ubiquitous in the industrial world and are of great use for monitoring of industrial assets. Most deep learning tools are designed for inputs of fixed and/or very limited size and many successful applications of deep learning to the industrial context use as inputs extracted features, which is a manually and often arduously obtained compact representation of the original signal. In this paper, we propose a fully unsupervised deep learning framework that is able to extract a meaningful and sparse representation of raw HF signals. We embed in our architecture impo…
École Polytechnique Fédérale de Lausanne (EPFL)（洛桑联邦理工学院）
Lipschitz-constrained neural networks have many applications in machine learning. Since designing and training expressive Lipschitz-constrained networks is very challenging, there is a need for improved methods and a better theoretical understanding. Unfortunately, it turns out that ReLU networks have provable disadvantages in this setting. Hence, we propose to use learnable spline activation functions with at least 3 linear regions instead. We prove that this choice is optimal among all component-wise $1$-Lipschitz activation functions in the sense that no other weight constrained architectur…
In recent years, deep neural networks (DNNs) achieved unprecedented performance in many low-level vision tasks. However, state-of-the-art results are typically achieved by very deep networks, which can reach tens of layers with tens of millions of parameters. To make DNNs implementable on platforms with limited resources, it is necessary to weaken the tradeoff between performance and efficiency. In this paper, we propose a new activation unit, which is particularly suitable for image restoration problems. In contrast to the widespread per-pixel activation units, like ReLUs and sigmoids, our un…
Deep neural networks (DNNs) generate much richer function spaces than shallow networks. Since the function spaces induced by shallow networks have several approximation theoretic drawbacks, this explains, however, not necessarily the success of deep networks. In this article we take another route by comparing the expressive power of DNNs with ReLU activation function to piecewise linear spline methods. We show that MARS (multivariate adaptive regression splines) is improper learnable by DNNs in the sense that for any given function that can be expressed as a function in MARS with $M$ parameter…
Technical University Munich（慕尼黑工业大学）；AI Research, Volkswagen Group（大众集团人工智能研究部）
We propose stochastic, non-parametric activation functions that are fully learnable and individual to each neuron. Complexity and the risk of overfitting are controlled by placing a Gaussian process prior over these functions. The result is the Gaussian process neuron, a probabilistic unit that can be used as the basic building block for probabilistic graphical models that resemble the structure of neural networks. The proposed model can intrinsically handle uncertainties in its inputs and self-estimate the confidence of its predictions. Using variational Bayesian inference and the central lim…
University of Illinois at Urbana Champaign（伊利诺伊大学厄巴纳-香槟分校）；Indian Institute of Technology Delhi（印度理工学院德里分校）
The scope of research in the domain of activation functions remains limited and centered around improving the ease of optimization or generalization quality of neural networks (NNs). However, to develop a deeper understanding of deep learning, it becomes important to look at the non linear component of NNs more carefully. In this paper, we aim to provide a generic form of activation function along with appropriate mathematical grounding so as to allow for insights into the working of NNs in future. We propose "Self-Learnable Activation Functions" (SLAF), which are learned during training and a…
National Yang Ming Chiao Tung University（阳明交通大学）；University at Albany - SUNY（奥尔巴尼大学 - 纽约州立大学）；University at Buffalo（布法罗大学）
Pooling is a simple but essential layer in modern deep CNN architectures for feature aggregation and extraction. Typical CNN design focuses on the conv layers and activation functions, while leaving the pooling layers with fewer options. We introduce the Learning Discrete Wavelet Pooling (LDW-Pooling) that can be applied universally to replace standard pooling operations to better extract features with improved accuracy and efficiency. Motivated from the wavelet theory, we adopt the low-pass (L) and high-pass (H) filters horizontally and vertically for pooling on a 2D feature map. Feature sign…
Microwave Remote Sensing Lab Center of Studies in Resources Engineering（微波遥感实验室 资源工程研究中心）；Centre of Machine Intelligence and Data Science（智能与数据科学中心）；Department of Electrical Engineering（电气工程系）；School of Mathematics and Statistics（数学与统计学学院）；Center of Studies in Resources Engineering（资源工程研究中心）
Graduate Institute of Electrical Engineering, National Taiwan University（台湾大学电机工程学研究所）
Recently, deep convolutional neural networks (CNNs) have achieved many eye-catching results. However, deploying CNNs on resource-constrained edge devices is constrained by limited memory bandwidth for transmitting large intermediated data during inference, i.e., activation. Existing research utilizes mixed-precision and dimension reduction to reduce computational complexity but pays less attention to its application for activation compression. To further exploit the redundancy in activation, we propose a learnable mixed-precision and dimension reduction co-design system, which separates channe…
TU Delft（代尔夫特理工大学）；Shell Global Solutions International B.V.（壳牌全球解决方案国际有限公司）
Binary Neural Networks (BNNs) are receiving an upsurge of attention for bringing power-hungry deep learning towards edge devices. The traditional wisdom in this space is to employ sign() for binarizing featuremaps. We argue and illustrate that sign() is a uniqueness bottleneck, limiting information propagation throughout the network. To alleviate this, we propose to dispense sign(), replacing it with a learnable activation binarizer (LAB), allowing the network to learn a fine-grained binarization kernel per layer - as opposed to global thresholding. LAB is a novel universal module that can sea…
Peking University（北京大学）；Nanjing University of Information Science and Technology（南京信息工程大学）
Recently, a number of learning-based optimization methods that combine data-driven architectures with the classical optimization algorithms have been proposed and explored, showing superior empirical performance in solving various ill-posed inverse problems, but there is still a scarcity of rigorous analysis about the convergence behaviors of learning-based optimization. In particular, most existing analyses are specific to unconstrained problems but cannot apply to the more general cases where some variables of interest are subject to certain constraints. In this paper, we propose Differentia…
University of Liverpool（利物浦大学）；Boston University（波士顿大学）；Northwestern University（西北大学）；University of Dayton（代顿大学）
We present POLAR, a polynomial arithmetic-based framework for efficient bounded-time reachability analysis of neural-network controlled systems (NNCSs). Existing approaches that leverage the standard Taylor Model (TM) arithmetic for approximating the neural-network controller cannot deal with non-differentiable activation functions and suffer from rapid explosion of the remainder when propagating the TMs. POLAR overcomes these shortcomings by integrating TM arithmetic with \textbf{Bernstein B{é}zier Form} and \textbf{symbolic remainder}. The former enables TM propagation across non-differentia…
Shandong University（山东大学）；CSIRO（澳大利亚联邦科学与工业研究组织）；Nanjing University of Science and Technology（南京理工大学）；Shandong Jianzhu University（山东建筑大学）
Non-linear activation functions, e.g., Sigmoid, ReLU, and Tanh, have achieved great success in neural networks (NNs). Due to the complex non-linear characteristic of samples, the objective of those activation functions is to project samples from their original feature space to a linear separable feature space. This phenomenon ignites our interest in exploring whether all features need to be transformed by all non-linear functions in current typical NNs, i.e., whether there exists a part of features arriving at the linear separable feature space in the intermediate layers, that does not require…
Subtractive Modulative Network with Learnable Periodic Activations
School of Electrical Engineering and Computer Science, KTH Royal Institute of Technology, Sweden（电气工程与计算机科学学院，皇家理工学院，瑞典）；School of Computer and Communication Sciences, EPFL, Switzerland（计算机与通信科学学院，瑞士联邦理工学院）
