Uncertainty

• Epistemic Uncertainity is the uncertainity due to lack of knowledge,

  In machine learning connection to Bayesian inference, an approach to model epistemic uncertainity is to draw samples from posterior predictive distribution which marginalizes over some posterior uncertainity about the model. In simple terms, it means that we can train/fit multiple models on the dataset and then draw samples from those models.

• Aliatoric Uncertainity (Aliator is latin for dice player) is the uncertainity due to inherent randomness