ICML
2005
IEEE
14 years 12 months 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...