ABGD parametrizes piecewise linear functions as difference of max-affine functions and converges linearly to an epsilon-accurate solution with O(d max(sigma/epsilon,1)^2) samples under sub-Gaussian noise, which is minimax optimal up to logs.
arXiv preprint arXiv:2207.09660 , year=
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
stat.ML 2verdicts
UNVERDICTED 2representative citing papers
Sp-GD recovers sparse max-affine parameters to epsilon accuracy with O(s log(d/s)) samples in the noise-free case under sub-Gaussian assumptions, supported by sparse-PCA initialization and an RMD transformation for generalized polynomials.
citing papers explorer
-
Locally Near Optimal Piecewise Linear Regression in High Dimensions via Difference of Max-Affine Functions
ABGD parametrizes piecewise linear functions as difference of max-affine functions and converges linearly to an epsilon-accurate solution with O(d max(sigma/epsilon,1)^2) samples under sub-Gaussian noise, which is minimax optimal up to logs.
-
Sparse Max-Affine Regression
Sp-GD recovers sparse max-affine parameters to epsilon accuracy with O(s log(d/s)) samples in the noise-free case under sub-Gaussian assumptions, supported by sparse-PCA initialization and an RMD transformation for generalized polynomials.