76
Voted
COLT
1993
Springer
15 years 1 months ago
1993
Springer
We introduce a new formal model in which a learning algorithm must combine a collection of potentially poor but statistically independent hypothesis functions in order to approxima...