Small-size epsilon-nets for axis-parallel rectangles and boxes.
Boris AronovEsther EzraMicha SharirPublished in: STOC (2009)
Keyphrases
- small size
- axis parallel rectangles
- multiple instance
- concept classes
- sample complexity
- membership queries
- learning theory
- multiple instance learning
- concept class
- learning problems
- special case
- vc dimension
- concept learning
- lower bound
- target concept
- real valued
- semi supervised learning
- learning algorithm
- regression problems
- multi modal
- efficient learning
- uniform distribution
- generalization error
- boolean functions
- theoretical analysis
- supervised learning
- active learning
- pairwise