Gaussian Processes and Kernel Methods¶
Kanagawa, M. (2001). Gaussian Processes and Kernel Methods: A Review on Connections and Equivalences.
Cited by¶
1 citation across 1 artifact.
Each citation links to the sentence it supports in the citing article.
Domain-specific¶
- Bayesian Interpretation of Kernel Regularization
- Finite-dimensional negative log densities often correspond to penalties, but infinite-dimensional Gaussian measures do not admit an ordinary Lebesgue density, and GP sample paths generally do not belong to the covariance kernel's RKHS when that RKHS is infinite-dimensional
This sourceThe measure-theoretic clauses about Lebesgue densities are background it does not supply. The review that states the equivalence explicitly for a general RKHS: its Proposition 3.6 gives the kernel-ridge estimator and the GP posterior mean as the same finite Gram-matrix solve under sigma-squared equal to n times lambda.
- Finite-dimensional negative log densities often correspond to penalties, but infinite-dimensional Gaussian measures do not admit an ordinary Lebesgue density, and GP sample paths generally do not belong to the covariance kernel's RKHS when that RKHS is infinite-dimensional
Verification¶
This reference passed the adversarial substantiation pipeline: it was checked to exist and to support the claim it is attached to. See how references were verified.
Registry ID ref:4ec8172bf194 · see in the full table