Product Unit Neural Networks with Constant Depth and Superlinear VC Dimension.
Michael SchmittPublished in: ICANN (2001)
Keyphrases
- vc dimension
- neural network
- upper bound
- lower bound
- concept classes
- sample complexity
- statistical learning theory
- covering numbers
- sample size
- distribution free
- inductive inference
- vapnik chervonenkis dimension
- empirical risk minimization
- vapnik chervonenkis
- mind change complexity
- pac learning
- generalization bounds
- concept class
- worst case
- compression scheme
- euclidean space
- pac learnability
- uniform convergence
- function classes
- neural network model
- sufficient conditions
- small number