Functional loops: monitoring functional organization of deep neural networks using algebraic topology

DOI10.1016/J.NEUNET.2024.106239MaRDI QIDQ6492701zbMATH OpenFDO

Authors Ben Zhang, Hongwei Lin

Publication date 25 April 2024

Published in Neural Networks (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1016/J.NEUNET.2024.106239

zbMATH Keywords

deep neural networkfunctional networktopological data analysis

Mathematics Subject Classification ID

Persistent homology and applications, topological data analysis (55N31) Artificial neural networks and deep learning (68T07)

Recommendations

• Topological approaches to deep learning
• Barcodes as summary of loss function topology
• 1-dimensional topological invariants to estimate loss surface non-convexity
• Quantification on the generalization performance of deep neural network with Tychonoff separation axioms
• Topological properties of the set of functions generated by neural networks of fixed size

Cites work

• Topology of deep neural networks

Cited in

(3)

• 1-dimensional topological invariants to estimate loss surface non-convexity
• Understanding the role of pathways in a deep neural network
• Quantification on the generalization performance of deep neural network with Tychonoff separation axioms

This page was built for publication: Functional loops: monitoring functional organization of deep neural networks using algebraic topology

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6492701)