Research breakthrough · curated entry
Backpropagation reaches Nature; multilayer networks become trainable
Learning representations by back-propagating errors
- Date
- October 9, 1986
- Category
- Research breakthrough
- Organization / Model
- Carnegie Mellon University / University of Toronto
- Evidence
- 1 source · 1 evidence family
Rumelhart, Hinton and Williams published the back-propagation algorithm in Nature, showing that multilayer networks can learn internal representations by propagating error backwards. The method became the foundation of how deep networks are trained today.
Curator's note: Cross-checked against the Nature article page.
Curated sources · 1
Cite this page
This page is a curated static snapshot and can be cited directly. Preferred format GB/T 7714—2025 (electronic resource [EB/OL]); the other 5 styles open below, together with the machine-readable files.
GB/T 7714—2025 (electronic resource [EB/OL])
AI 纪年. 反向传播论文登上《Nature》，多层网络变得可训练[EB/OL]. (1986-10-09)[2026-11-03]. https://aihist.net/e/backprop-1986.html.
APA 7 · American Psychological Association
AI 纪年. (1986, October 9). 反向传播论文登上《Nature》，多层网络变得可训练 [Web page]. https://aihist.net/e/backprop-1986.html
MLA 9 · Modern Language Association
"反向传播论文登上《Nature》，多层网络变得可训练." AI 纪年, 9 Oct. 1986, aihist.net/e/backprop-1986.html. Accessed 3 Nov. 2026.
Chicago 17 · notes-bibliography
AI 纪年. "反向传播论文登上《Nature》，多层网络变得可训练." Published Oct. 9, 1986. Accessed Nov. 3, 2026. https://aihist.net/e/backprop-1986.html.
IEEE · citation-reference list
"反向传播论文登上《Nature》，多层网络变得可训练," AI 纪年, 9 Oct. 1986. [Online]. Available: https://aihist.net/e/backprop-1986.html. [Accessed: 3 Nov. 2026].
ACM · author-date variant
AI 纪年. October 9, 1986. 反向传播论文登上《Nature》，多层网络变得可训练. Retrieved November 3, 2026 from https://aihist.net/e/backprop-1986.html.
All six styles come from one template and can be compared word for word. [引用日期], Accessed, urldate and Y2 all carry this page's build date, 2026-11-03 — a static page is an immutable snapshot, so substitute the date you actually consulted it. The address in every style is the archive's canonical page: this English page translates the same snapshot, and the work keeps its name, so one record cites both. For one-click copying use the same entry in the timeline, which takes the day you opened it. Expand the other styles before printing.