Ground truth
Referans gerçeklik
D4
Ground truth is task-specific reference information or an accepted target used to compare a system's output.
Review status: 2026-11-26
Technical explanation
It gives an evaluation a reference for comparison, calculation, or judgment.
Conceptual boundaries
Ground truth is not infallible truth; human annotation can contain ambiguity and disagreement.
Provider-neutral example
For a delivery-status task, a reviewed record stating that an order was delivered can be the reference used to compare a model's classification.
Limitations
Ambiguity and disagreement can remain in task-specific reference information and affect how comparison results are interpreted.
Related concepts
Atomic claims and evidence
1.1NIST AI 600-1, Generative Artificial Intelligence Profile — D4 evidence slice
- Source
- NIST AI 600-1, Generative Artificial Intelligence Profile — D4 evidence slice
- Source role
- Authoritative source
- Exact locator
- MAP 2.3 action text, printed p. 27
- Supported claim
- NIST lists comparison with known ground-truth data as one evaluation method for examining AI-system information integrity.
- Last verification
- Review due
- Scope limitation
- The source presents ground truth as one method among several, not as a universal solution for every evaluation target.
2.1NIST AI 100-1, Artificial Intelligence Risk Management Framework (AI RMF 1.0) — D4 evidence slice
- Source
- NIST AI 100-1, Artificial Intelligence Risk Management Framework (AI RMF 1.0) — D4 evidence slice
- Source role
- Authoritative source
- Exact locator
- Section 3, printed p. 12
- Supported claim
- NIST says that human judgment should be used when deciding specific metrics and threshold values for AI trustworthiness characteristics.
- Last verification
- Review due
- Scope limitation
- This supports the need to make judgment explicit; it does not prescribe one labelling method or eliminate disagreement.
3.1Aroyo and Welty, Truth Is a Lie: Crowd Truth and the Seven Myths of Human Annotation
- Source
- Aroyo and Welty, Truth Is a Lie: Crowd Truth and the Seven Myths of Human Annotation
- Source role
- Authoritative source
- Exact locator
- Abstract; "The Seven Myths" and "Disagreement Is Bad," printed pp. 16-17
- Supported claim
- Aroyo and Welty describe human annotation as a process in which ambiguity and disagreement can carry information rather than simply being discarded as error.
- Last verification
- Review due
- Scope limitation
- This is an account of human annotation and does not say that every disagreement is desirable or that every task lacks a usable reference.