Nearly-tight VC-dimension and Pseudodimension Bounds for Piecewise Linear Neural Networks.
Peter L. BartlettNick HarveyChristopher LiawAbbas MehrabianPublished in: J. Mach. Learn. Res. (2019)
Keyphrases
- vc dimension
- piecewise linear
- upper bound
- lower bound
- neural network
- worst case
- vapnik chervonenkis
- generalization bounds
- covering numbers
- sample complexity
- distribution free
- concept classes
- dynamic programming
- statistical learning theory
- sample size
- empirical risk minimization
- inductive inference
- uniform convergence
- hyperplane
- pattern recognition
- function classes
- compression scheme
- concept class
- np hard
- data sets
- image processing
- learning algorithm
- objective function
- special case