145
click to vote
NIPS
2007
15 years 4 months ago
2007
Empirical risk minimization offers well-known learning guarantees when training and test data come from the same domain. In the real world, though, we often wish to adapt a classi...