Heishiro Kanagawa
Heishiro Kanagawa
Home
Publications
Contact
Light
Dark
Automatic
Bayesian computation
Controlling Moments with Kernel Stein Discrepancies
The kernel Stein discrepancy (KSD) is a measure of probabilistic discrepancy, which can be used to evaluate the quality of Monte Carlo samplers. It is shown that the KSD controls moments with an appropriate kernel.
Heishiro Kanagawa (金川 平志郎)
PDF
Code
ArXiv
Cite
×