Training a Single Sigmoidal Neuron Is Hard

DOI10.1162/089976602760408035MaRDI QIDQ4409384zbMATH OpenOpenAlexDBLPWikidataFDO

Authors Jiří Šíma

Publication date 14 October 2003

Published in Neural Computation (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1162/089976602760408035

zbMATH Keywords

constructive learningfeedforward neural networkshardness results

Mathematics Subject Classification ID

Learning and adaptive systems in artificial intelligence (68T05)

Recommendations

• scientific article; zbMATH DE number 2080651
• The computational intractability of training sigmoidal neural networks
• scientific article; zbMATH DE number 2080455
• Computational limitations on training sigmoid neural networks
• Hardness results for neural network approximation problems

Cites work

• Computational limitations on learning from examples
• Computational limitations on training sigmoid neural networks
• Feedforward nets for interpolation and classification
• Learnability and the Vapnik-Chervonenkis dimension
• Learning representations by back-propagating errors
• On the complexity of loading shallow neural networks
• On the complexity of polyhedral separability
• On the geometric separability of Boolean functions
• Robust trainability of single neurons
• The computational intractability of training sigmoidal neural networks
• The densest hemisphere problem
• The hardness of approximate optima in lattices, codes, and systems of linear equations

Cited in

(18)

• Recovery of regular ridge functions on the ball
• Global optimization issues in deep network regression: an overview
• Complexity of training ReLU neural network
• On the Nonlearnability of a Single Spiking Neuron
• Long-Range Out-of-Sample Properties of Autoregressive Neural Networks
• The computational intractability of training sigmoidal neural networks
• scientific article; zbMATH DE number 2080455 (Why is no real title available?)
• scientific article; zbMATH DE number 2080651 (Why is no real title available?)
• On the infeasibility of training neural networks with small mean-squared error
• Neural networks using a logistics sigmoid function: linear classifier bounds and global nonattainability
• scientific article; zbMATH DE number 7626778 (Why is no real title available?)
• fPINNs: Fractional Physics-Informed Neural Networks
• Loading Deep Networks Is Hard: The Pyramidal Case
• Black holes and the loss landscape in machine learning
• Complexity of network training for classes of Neural Networks
• First Steps Towards a Runtime Analysis of Neuroevolution
• On the omnipresence of spurious local minima in certain neural network training problems
• Computational limitations on training sigmoid neural networks

This page was built for publication: Training a Single Sigmoidal Neuron Is Hard

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4409384)