Dimension-free convergence rates for gradient Langevin dynamics in RKHS.

Boris Muzellec Kanji Sato Mathurin Massias Taiji Suzuki

Published in: CoRR (2020)

Keyphrases

convergence rate
gradient method
gaussian kernels
reproducing kernel hilbert space
convergence speed
step size
learning rate
global convergence
conjugate gradient
kernel methods
dynamical systems
random projections
stopping criterion
loss function
primal dual
pairwise
learning algorithm
least squares
data points