Private PAC learning implies finite Littlestone dimension.
Noga AlonRoi LivniMaryanthe MalliarisShay MoranPublished in: CoRR (2018)
Keyphrases
- pac learning
- concept classes
- mistake bound
- sample complexity
- uniform distribution
- computational learning theory
- learning theory
- learning problems
- concept class
- sample size
- pac model
- decision lists
- vc dimension
- upper bound
- theoretical analysis
- target concept
- special case
- lower bound
- membership queries
- linear threshold
- learning algorithm
- supervised learning
- active learning
- generalization error
- data sets
- learning tasks
- training examples
- training set
- agnostic learning