Ground truth
Ground truth is reference information treated as correct for assessing the accuracy or correctness of predictions, labels, or other outputs.
Human-in-the-Loop Machine Learning: Active learning and annotation for human-centered AI
Reference labels or outputs treated as correct for a data item.
- How it is used
- Uses known answers to measure annotation quality and support evaluation of human review tasks.
- How deep it goes
- Explains sampling strategies, expert creation, error margins, and bounding-box comparison, including risks from flawed reference data.
- Examples
- Illustrates accuracy with 80 correct known answers out of 100 and expert-reviewed bounding boxes.
- What you can do after
- Evaluate annotation quality by choosing reference samples and comparing responses with known answers.
Evaluating AI Systems
Ground truth is the reference standard used to judge whether system outputs are correct, acceptable, or incorrect.
- How it is used
- Establishes a shared reference for judging outputs when defining an AI task for evaluation.
- How deep it goes
- Distinguishes criteria-based and example-based definitions while deferring detailed judgment rubrics and annotation guidelines.
- Examples
- Illustrates the standard with medical chatbot criteria and patient questions paired with acceptable and unacceptable responses.
- What you can do after
- Evaluate and understand whether a task’s initial ground truth resolves fundamental disagreements without specifying every edge case.
Also covered in
Related concepts
- inter-rater reliabilityused by · Inter-rater reliability is the degree to which independent raters give consistent judgments or labels to the same items, subjects, or events.
- accuracymeasures · Accuracy is the degree to which a result agrees with a specified truth, reference, or criterion, commonly measured as the proportion of evaluated predictions that are correct.
- confusion matrixused by · A confusion matrix is a table that counts classification outcomes by comparing predicted classes with actual classes.
- factualitymeasured by · Factuality is the degree to which claims in content accurately represent externally verifiable facts.
Questions about ground truth
What is ground truth?
Ground truth is reference information treated as correct for assessing the accuracy or correctness of predictions, labels, or other outputs.
Which Manning books explain ground truth?
Human-in-the-Loop Machine Learning: Active learning and annotation for human-centered AI and Evaluating AI Systems.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.

