Sparse fourier transform in any constant dimension with nearly-optimal sample complexity in sublinear time.
Michael KapralovPublished in: STOC (2016)
Keyphrases
- fourier transform
- sample complexity
- frequency domain
- constant factor
- fourier coefficients
- learning problems
- vc dimension
- theoretical analysis
- signal processing
- pac learning
- supervised learning
- lower bound
- upper bound
- active learning
- fourier domain
- discrete fourier transform
- special case
- generalization error
- concept classes
- optimal solution
- dynamic programming
- learning algorithm
- semi supervised
- worst case
- pairwise
- image segmentation
- neural network
- image processing