Overcoming Catastrophic Forgetting by Bayesian Generative Regularization
Patrick H. Chen 1 Wei Wei 2 Cho-jui, Hsieh 1 Bo Dai 3
Abstract els fail to tackle continual learning with the phenomenon of
catastrophic forgetting, where deep neural networks tend
The streaming update of Bayesian posterior cal-
to lose the information of previous tasks (i.e. classification
culation pro ...


雷达卡




京公网安备 11010802022788号







