English
Related papers

Related papers: Can an MLP Absorb Its Own Skip Connection?

200 papers

Rectified Linear Units (ReLU) are the default choice for activation functions in deep neural networks. While they demonstrate excellent empirical performance, ReLU activations can fall victim to the dead neuron problem. In these cases, the…

Machine Learning · Computer Science 2023-02-14 Tim Whitaker , Darrell Whitley

We propose and analyze a new family of algorithms for training neural networks with ReLU activations. Our algorithms are based on the technique of alternating minimization: estimating the activation patterns of each ReLU for all given…

Machine Learning · Computer Science 2018-10-12 Gauri Jagatap , Chinmay Hegde

Deep neural networks, particularly those employing Rectified Linear Units (ReLU), are often perceived as complex, high-dimensional, non-linear systems. This complexity poses a significant challenge to understanding their internal learning…

Machine Learning · Computer Science 2025-11-11 Longqing Ye

Activation functions are fundamental to deep neural networks, governing gradient flow, optimization stability, and representational capacity. Within historic deep architectures, while ReLU has been the dominant choice for the activation…

Machine Learning · Computer Science 2026-03-10 Mingi Kang , Zai Yang , Jeova Farias Sales Rocha Neto

A key challenge in modern deep learning theory is to explain the remarkable success of gradient-based optimization methods when training large-scale, complex deep neural networks. Though linear convergence of such methods has been proved…

Machine Learning · Computer Science 2025-09-30 Yash Jakhmola

We study the problem of inverting a deep generative model with ReLU activations. Inversion corresponds to finding a latent code vector that explains observed measurements as much as possible. In most prior works this is performed by…

Machine Learning · Computer Science 2019-06-20 Qi Lei , Ajil Jalal , Inderjit S. Dhillon , Alexandros G. Dimakis

Nonlinear activation functions are widely recognized for enhancing the expressivity of neural networks, which is the primary reason for their widespread implementation. In this work, we focus on ReLU activation and reveal a novel and…

Machine Learning · Computer Science 2025-10-22 Chaoyue Liu , Han Bi , Like Hui , Xiao Liu

Despite their empirical success, pushing Transformer architectures to extreme depth often leads to a paradoxical failure: representations become increasingly redundant, lose rank, and ultimately collapse. Existing explanations largely…

Machine Learning · Computer Science 2026-01-16 Haoran Su , Chenyu You

In this paper, we prove that a shallow neural network with a monotone sigmoid, ReLU, ELU, Softplus, or LeakyReLU activation function can arbitrarily well approximate any L^p(p>=2) integrable functions defined on R*[0,1]^n. We also prove…

Machine Learning · Computer Science 2021-10-12 Ming-Xi Wang , Yang Qu

We consider the embedding of piecewise-linear deep neural networks (ReLU networks) as surrogate models in mixed-integer linear programming (MILP) problems. A MILP formulation of ReLU networks has recently been applied by many authors to…

Optimization and Control · Mathematics 2022-01-10 Bjarne Grimstad , Henrik Andersson

One of the mysteries in the success of neural networks is randomly initialized first order methods like gradient descent can achieve zero training loss even though the objective function is non-convex and non-smooth. This paper demystifies…

Machine Learning · Computer Science 2019-02-06 Simon S. Du , Xiyu Zhai , Barnabas Poczos , Aarti Singh

Gradient descent (GD) type optimization schemes are the standard instruments to train fully connected feedforward artificial neural networks (ANNs) with rectified linear unit (ReLU) activation and can be considered as temporal…

Optimization and Control · Mathematics 2022-02-24 Shokhrukh Ibragimov , Arnulf Jentzen , Timo Kröger , Adrian Riekert

Despite remarkable performance on a variety of tasks, many properties of deep neural networks are not yet theoretically understood. One such mystery is the depth degeneracy phenomenon: the deeper you make your network, the closer your…

Machine Learning · Statistics 2025-11-18 Cameron Jakub , Mihai Nica

Transformers often fail to learn generalizable algorithms, instead relying on brittle heuristics. Using graph connectivity as a testbed, we explain this phenomenon both theoretically and empirically. We consider a simplified Transformer…

Machine Learning · Computer Science 2026-02-19 Qilin Ye , Deqing Fu , Robin Jia , Vatsal Sharan

In many recent works, multi-layer perceptions (MLPs) have been shown to be suitable for modeling complex spatially-varying functions including images and 3D scenes. Although the MLPs are able to represent complex scenes with unprecedented…

Computer Vision and Pattern Recognition · Computer Science 2023-07-04 Animesh Karnewar , Tobias Ritschel , Oliver Wang , Niloy J. Mitra

In recent work it has been shown that determining a feedforward ReLU neural network to within high uniform accuracy from point samples suffers from the curse of dimensionality in terms of the number of samples needed. As a consequence,…

Machine Learning · Statistics 2024-11-11 Yuanyuan Li , Philipp Grohs , Philipp Petersen

In this work, we demonstrate provable guarantees on the training of a single ReLU gate in hitherto unexplored regimes. We give a simple iterative stochastic algorithm that can train a ReLU gate in the realizable setting in linear time while…

Machine Learning · Computer Science 2022-04-04 Sayar Karmakar , Anirbit Mukherjee

We here study whether training biases can make hidden neurons specialize in minimal one-hidden-layer MLPs, and whether such specialization improves prototype-based reconstruction of the training dataset from the learned weights. We consider…

Machine Learning · Computer Science 2026-05-26 Enrique Alba , Ezequiel Lopez-Rubio

ReLU neural networks trained as surrogate models can be embedded exactly in mixed-integer linear programs (MILPs), enabling global optimization over the learned function. The tractability of the resulting MILP depends on structural…

Optimization and Control · Mathematics 2026-04-27 Calvin Tsay

In this paper, it is demonstrated through a case study that multilayer feedforward neural networks activated by ReLU functions can in principle be trained iteratively with Mixed Integer Linear Programs (MILPs) as follows. Weights are…

Machine Learning · Computer Science 2021-04-12 Steffen Goebbels
‹ Prev 1 3 4 5 6 7 10 Next ›