gaussian process

left off at ### Kernels

The goal of a gaussian process is to model a multivariate gaussian distribution from training data.
The joint probability distribution $P_{X,Y}$ of the training data $Y$ and the test data $X$ spans the space of possible function values for the function we want to predict. It has $|X|+|Y|$ dimensions

$$k:{\mathbb{R}}^n\times {\mathbb{R}}^n\to \mathbb{R},\Sigma =\text{Cov}(X,X^{\prime })=k(t,t^{\prime })$$

References

https://distill.pub/2019/visual-exploration-gaussian-processes/

regression
multivariate gaussian distribution