Towards Lower Bounds on the Depth of ReLU Neural Networks

DOI10.1137/22m1489332zbMath1529.68276arXiv2105.14835OpenAlexW3211910768MaRDI QIDQ6100606

Amitabh Basu, Christoph Hertrich, Martin Skutella, Marco Di Summa

Publication date: 22 June 2023

Published in: SIAM Journal on Discrete Mathematics (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/2105.14835

zbMATH Keywords

expressivitymixed-integer optimizationpolyhedral methodsReLU neural networksdepth bounds

Mathematics Subject Classification ID

Artificial neural networks and deep learning (68T07) Applications of mathematical programming (90C90) Mixed integer programming (90C11) Computational aspects related to convexity (52B55)

Cites Work

• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Approximation and estimation bounds for artificial neural networks
• Learning a classification of mixed-integer quadratic programming problems
• Degree of approximation by neural and translation networks with a single hidden layer
• Learning when to use a decomposition
• Machine learning for combinatorial optimization: a methodological tour d'horizon
• Lossless compression of deep neural networks
• Approximation spaces of deep neural networks
• Neural networks with linear threshold activations: structure and algorithms
• Error bounds for approximations with deep ReLU networks
• On learning and branching: a survey
• Complexity of training ReLU neural network
• Virtual polytopes
• Relu Deep Neural Networks and Linear Finite Elements
• Generalization of Hinging Hyperplanes
• Universal approximation bounds for superpositions of a sigmoidal function
• General Constructive Representations for Continuous Piecewise-Linear Functions
• Neural Network Learning
• The Computational Complexity of ReLU Network Training Parameterized by Data Dimensionality
• Essentials of Tropical Combinatorics
• Approximation Algorithms for Training One-Node ReLU Neural Networks
• A Machine Learning-Based Approximation of Strong Branching
• Sharp Bounds for the Number of Regions of Maxout Networks and Vertices of Minkowski Sums
• Approximation by superpositions of a sigmoidal function
• Strong mixed-integer programming formulations for trained neural networks
• ReLU neural networks of polynomial size for exact maximum flow computation

This page was built for publication: Towards Lower Bounds on the Depth of ReLU Neural Networks