Claim · 2018 · publication · Stated in source · Supported

According to Sutton and Barto (2018), the off-policy temporal-difference control algorithm known as Q-learning is due to Watkins (1989).

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

Reinforcement Learning: An Introduction (second edition) → attributes to → Learning from Delayed Rewards

Where it appears

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

Claim and evidence

Reinforcement Learning: An Introduction (second edition)attributes to ↓Learning from Delayed Rewards

Machine-checked, not reviewed by an editorStated in sourceSupported

Supporting evidence: primary · 1 source

Date: 2018 · publication Records dated 2018 →
Kind of attribution: Attributes a term or concept to the work

Scope and limitsAttributed statement; "breakthroughs" is the book's word.

The evidence 02

primary sourceSupports the claim

Reinforcement Learning: An Introduction (second edition)

MIT Press; author-hosted PDF (incompleteideas.net) · 2018

Cite the original source

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

Source excerpt

“One of the early breakthroughs in reinforcement learning was the development of an … TD control algorithm known as Q-learning (Watkins, 1989)”

Location in source PDF p. 153, section 6.5

How it was read Retrieved directly from the publisher

Machine-checked · 2026-10-31

Source details and check notes

Authors: Richard S. Sutton; Andrew G. Barto
Version: Retrieved for the candidate pack on 2026-10-29 (SHA-256 2dd0d71d9ee883fbeb99f9b888c65ac3255fae2512203d11b18e838befffe9a6). Author-hosted PDF of the second edition, copyright 2018, 2020 (CC BY-NC-ND 2.0). Ligatures such as "ff" are mis-encoded in the text layer; quotations avoid them.
Canonical address: Open the canonical source ↗
Accessed: 2026-10-29
Language: English (en)
Attribution: Richard S. Sutton, Andrew G. Barto, “Reinforcement Learning: An Introduction (second edition)”, 2018.
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 quotation (21 words) matched the text retrieved for the candidate pack on 2026-10-29. Matching ignored differences of whitespace, case, quotation marks, dashes, PDF line-break hyphens and Unicode compatibility forms, so the source’s typography can differ. The elided word is "off-policy", mis-encoded in the text layer. The source role is primary for this record: the passage is the citing text’s own attribution, which is what the record states (corrected at admission; the pack had secondary, relative to the work credited). A cross-family machine check (31 October 2026, item 75) re-read the record’s sources and found that they support what the record states. No human editorial review.

primary sourceSupports the claim

Reinforcement Learning: An Introduction (second edition)

MIT Press; author-hosted PDF (incompleteideas.net) · 2018

Cite the original source

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

Source excerpt

“The most influential integration of reinforcement learning and MDPs is due to Watkins (1989).”

Location in source PDF p. 92, chapter 3 notes

How it was read Retrieved directly from the publisher

Machine-checked · 2026-10-31

Source details and check notes

Authors: Richard S. Sutton; Andrew G. Barto
Version: Retrieved for the candidate pack on 2026-10-29 (SHA-256 2dd0d71d9ee883fbeb99f9b888c65ac3255fae2512203d11b18e838befffe9a6). Author-hosted PDF of the second edition, copyright 2018, 2020 (CC BY-NC-ND 2.0). Ligatures such as "ff" are mis-encoded in the text layer; quotations avoid them.
Canonical address: Open the canonical source ↗
Accessed: 2026-10-29
Language: English (en)
Attribution: Richard S. Sutton, Andrew G. Barto, “Reinforcement Learning: An Introduction (second edition)”, 2018.
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 quotation (14 words) matched the text retrieved for the candidate pack on 2026-10-29. Matching ignored differences of whitespace, case, quotation marks, dashes, PDF line-break hyphens and Unicode compatibility forms, so the source’s typography can differ. The source role is primary for this record: the passage is the citing text’s own attribution, which is what the record states (corrected at admission; the pack had secondary, relative to the work credited). A cross-family machine check (31 October 2026, item 75) re-read the record’s sources and found that they support what the record states. 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. A document expressly attributes an argument, concept, term, finding or approach to another work, in its own words; the kind of attribution is recorded. The record is the citing document’s claim. It is not evidence that the cited work originated the idea, that the attribution is accurate, or of influence beyond the passage. Citation alone is recorded as a citation.

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

Record ID th-rl-sb-credits-watkins