Summary Part 2: Deep Gaussian processes

• A Gaussian distribution depends on a mean and a covariance vector / matrix. • A Gaussian process depends on a mean and a covariance function. Next: Demo, from Gaussian distributions to Gaussian processes. Infinite model... but we always work with finite sets! p(fA, fB) ∼ N (µ,K). with: µ = µA µB and K