👉 In computational complexity theory, an "epoch" is a data structure that represents a time interval in a machine learning model. An epoch is used to track how long it takes for a model's predictions to change from one epoch to another. This concept is important in understanding the trade-off between prediction speed and precision. An epoch can be defined as the time taken by a model to make its predictions on new data points, or equivalently, the number of epochs required to reach a certain accuracy threshold.