Uncertainty in AI may arise from the data (e.g. inaccuracy of {[sensors}}) or due to the processing of the data (e.g. stochastic nature of some neural-network training methods). In uncertainty may affect both the training sets used to create machine learning or other forms of AI; and the data needed for decsion maing when the AI model is deployed. For example, if an image recognistion algorithm is trained using past radiographic images, some may have been misclassified by previous human processes (uncertaunty in training data), but also when deployed some scans may be poor quality due to patients miving during scanning. There are various methods used in AI for uncertain reasoning. In addiiton, it is important when designing the human processes and user interactions that unceertainty is adequately communicated and overall processes created that are {[robust}}.
Used on Chap. 3: pages 40, 45, 51; Chap. 7: page 147; Chap. 11: pages 221, 239; Chap. 15: page 355; Chap. 18: pages 429, 430, 440, 448, 451; Chap. 19: pages 456, 463, 472, 473, 478, 481