On PDE Characterization of Smooth Hierarchical Functions Computed by Neural Networks

DOI10.1162/NECO_A_01441OpenAlexW3209399971WikidataQ114066577 ScholiaQ114066577MaRDI QIDQ5034481

Konrad P. Körding, Khashayar Filom, Roozbeh Farhoodi

Publication date: 18 February 2022

Published in: Neural Computation (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/2005.08859

Mathematics Subject Classification ID

Computer science (68-XX)

Related Items (1)

Hyper-flexible convolutional neural networks based on generalized Lehmer and power means

Cites Work

• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Approximate complexity and functional representation
• The solutions of a smooth PDE can be dense in C[I]
• Why does deep and cheap learning work so well?
• Multilayer feedforward networks are universal approximators
• Topological properties of the set of functions generated by neural networks of fixed size
• A universal differential equation
• Characterization of Classes of Functions
• Theoretical Insights Into the Optimization Landscape of Over-Parameterized Shallow Neural Networks
• On Hilbert's thirteenth problem and related questions
• Spurious Valleys in Two-layer Neural Network Optimization Landscapes
• On Functions Computed on Trees
• On the Structure of Continuous Functions of Several Variables
• Approximation by superpositions of a sigmoidal function

This page was built for publication: On PDE Characterization of Smooth Hierarchical Functions Computed by Neural Networks