ground truth

Terms from Artificial Intelligence: humans at the heart of algorithms

The glossary is being gradually proof checked, but currently has many typos and misspellings.

In order to perform supervised learning, we need some form of ground truth, that is for at least some training data we need to know the correct classification or output value. This may be obtained through human labelling of data, from sensor data or in some cases simulation.

Used in Chap. 8: page 106; Chap. 12: page 183; Chap. 15: page 231; Chap. 18: pages 280, 281, 289