ICML
2005
IEEE
15 years 13 days ago
2005
IEEE
We provide a general framework for learning precise, compact, and fast representations of the Bayesian predictive distribution for a model. This framework is based on minimizing t...