Claim · 1986-10-09 · publication · Stated in source · Supported

Rumelhart, Hinton and Williams (Nature, 1986) describe back-propagation as a learning procedure for networks of neurone-like units that repeatedly adjusts connection weights to minimise the difference between actual and desired output, so that hidden units come to represent features of the task domain.

One recorded claim. The claim panel gives its qualification, the source passages with their place in each source, and how it was checked.

What it connects

Learning representations by back-propagating errors → discusses →

Where it appears

No collection page holds this claim. It is listed with its records in the explorer.

Claim and evidence

Learning representations by back-propagating errorsdiscusses ↓

Machine-checked, not reviewed by an editorStated in sourceSupported

Supporting evidence: primary · 1 source

Date: 1986-10-09 · publication Records dated 1986 →

Scope and limitsThe letter’s own description, which calls the procedure “new”; that is the authors’ statement, not a priority finding. Other texts’ attributions of backpropagation to earlier work are recorded separately.

The evidence 01

primary sourceSupports the claim

Learning representations by back-propagating errors

Nature 323; scan hosted on Geoffrey Hinton's University of Toronto page · 1986-10-09

Cite the original source

Cite the original document, not Selekoa. Check these details against it.

Paraphrase

The abstract says the authors describe a new learning procedure, back-propagation, for networks of neurone-like units; it repeatedly adjusts the weights of the connections to minimise a measure of the difference between the actual and desired output vectors, and as a result internal "hidden" units come to represent important features of the task domain.

Location in source PDF p. 1 (printed p. 533), abstract

How it was read Retrieved directly from the publisher

Machine-checked · 2026-10-31

Source details and check notes

Authors: David E. Rumelhart; Geoffrey E. Hinton; Ronald J. Williams
Version: Retrieved for the candidate pack on 2026-10-29 (SHA-256 d26997baf588222109d32545604a2a2ed400dc769a21fd49a5acdc4a955396ae). Scan of the Nature pages 533-536 (no text layer), author-hosted; printed pp. 533 and 536 read from page images.
Canonical address: Open the canonical source ↗
Accessed: 2026-10-29
Language: English (en)
Attribution: David E. Rumelhart, Geoffrey E. Hinton, Ronald J. Williams, “Learning representations by back-propagating errors”, 1986-10-09.
Reuse: No open redistribution licence asserted. Bibliographic metadata, editorial paraphrases and short attributed quotations (at most 50 words each). Recheck terms before wider redistribution.
Check notes: Candidate pack passage, Kinds of AI and technical history (research/technical-history). The paraphrase is Selekoa’s, written from the text retrieved for the candidate pack on 2026-10-29; it is not a quotation and was not machine-matched. Scanned PDF without a text layer: paraphrase of the page image. No human editorial review.

Status and meaning

Editorial status. Not reviewed by an editor. An exploratory record that no editor has accepted. This status does not say whether a person has read any linked passage. Each evidence record above shows how it was checked.

Relation. The concept is explicitly discussed in the document. This makes no earliest-use, invention or transmission claim.

Date. Date of publication, as the source gives it.

Record ID th-cn-def-rhw-1986