Learning representations by back-propagating errors
David E. Rumelhart, Geoffrey E. Hinton and Ronald J. Williams published Learning representations by back-propagating errors in Nature on 9 October 1986, describing a learning procedure that repeatedly adjusts connection weights in networks of neuron-like units to reduce the difference between actual and desired outputs.
Historical classification
This verified record belongs to LINEAiGE's artificial-intelligence history layer beginning in 1956.
Why it matters
A documented major milestone in the historical development and dissemination of back-propagation for training multilayer neural networks.
People, organizations & technologies
PEOPLE
David E. Rumelhart, Geoffrey E. Hinton, Ronald J. Williams
ORGANIZATIONS
University of California, San Diego, Carnegie-Mellon University, Nature
TECHNOLOGIES
Back-propagation, Neural networks
Evidence & verification
STATE
verified
CONFIDENCE
high
LAST VERIFIED
2026-09-07T22:49:13+08:00
Title, authors, publication date, volume, pages, DOI and the paper's description of its learning procedure are verified against Nature's record for the original article. This record does not claim that the authors were the first people ever to formulate back-propagation.
Sources
Learning representations by back-propagating errors
Nature
PUBLISHED
1986-10-09
Trace the chronology
Record identity
LINEAiGE IDbackpropagation-1986