Deep, big, simple neural nets for handwritten digit recognition

Dan Claudiu Cireşan, Ueli Meier, Luca Maria Gambardella, Jürgen Schmidhuber

Research output: Contribution to journalArticlepeer-review

628 Scopus citations

Abstract

Good old online backpropagation for plain multilayer perceptrons yields a very low 0.35% error rate on the MNIST handwritten digits benchmark. All we need to achieve this best result so far are many hidden layers, many neurons per layer, numerous deformed training images to avoid overfitting, and graphics cards to greatly speed up learning. © 2010 Massachusetts Institute of Technology.
Original languageEnglish (US)
Pages (from-to)3207-3220
Number of pages14
JournalNeural Computation
Volume22
Issue number12
DOIs
StatePublished - Dec 1 2010
Externally publishedYes

ASJC Scopus subject areas

  • Cognitive Neuroscience

Fingerprint

Dive into the research topics of 'Deep, big, simple neural nets for handwritten digit recognition'. Together they form a unique fingerprint.

Cite this