A case where a spindly two-layer linear network decisively outperforms any neural network with a fully connected input layer

Manfred K. Warmuth; Wojciech Kotłowski; Ehsan Amid

Scientific Information System of the Poznań University of Technology

PL EN

Main page / Publications / A case where a spindly two-layer linear network decisively outperforms any neural network with a fully connected input layer

Submit a comment

Chapter

Download BibTeX

Title

A case where a spindly two-layer linear network decisively outperforms any neural network with a fully connected input layer

Authors

Manfred K. Warmuth
Wojciech Kotłowski (WIiT) ^{[ 1 ][ 2.3 ][ P ]}
Ehsan Amid

^{[ 1 ]} Instytut Informatyki, Wydział Informatyki i Telekomunikacji, Politechnika Poznańska | ^{[ P ]} employee

Scientific discipline (Law 2.0)

[2.3] Information and communication technology

Year of publication

2021

Chapter type

chapter in monograph / paper

Publication language

english

Keywords

EN

neural networks
scale invariance
spindly network
gradient descent
lower bounds

Pages (from - to)

1214 - 1236

URL

http://proceedings.mlr.press/v132/warmuth21a.html

Book

Proceedings of the 32nd International Conference on Algorithmic Learning Theory

Presented on

32nd International Conference on Algorithmic Learning Theory (ALT 2021), 16-19.03.2021

Open Access Mode

publisher's website

Open Access Text Version

final published version

Date of Open Access to the publication

at the time of publication

Ministry points / chapter

5

Ministry points / conference (CORE)

70

System created by Poznań University of Technology and Poznan Supercomputing and Networking Center

Log in through eKonto to add to SIS