Monte Carlo Dropout: A technique used to estimate the uncertainty of a model, by randomly dropping out neurons during the forward pass and averaging the predictions over multiple forward passes.