83
Voted
AAAI
2007
15 years 23 days ago
2007
In semi-supervised learning, a number of labeled examples are usually required for training an initial weakly useful predictor which is in turn used for exploiting the unlabeled e...