Dual VC Dimension Obstructs Sample Compression by Embeddings.
Zachary ChaseBogdan ChornomazSteve HannekeShay MoranAmir YehudayoffPublished in: COLT (2024)
Keyphrases
- vc dimension
- sample size
- compression scheme
- euclidean space
- upper bound
- sample complexity
- vapnik chervonenkis dimension
- image compression
- covering numbers
- statistical learning theory
- concept classes
- inductive inference
- empirical risk minimization
- pac learning
- distribution free
- data compression
- lower bound
- compression ratio
- generalization bounds
- uniform convergence
- compression algorithm
- model selection
- pac learnability
- vector space
- concept class
- worst case
- low dimensional
- function classes
- high dimensional
- distance measure
- dimensionality reduction
- image processing