Bayesian Unlearning

An interesting problem in Dynamic mixture model is to be able to see how we can accommodate unlearning and if it can be done by a 'particle-filtering' like on-line approach