§31.1: by linearity of expectation the generalization loss of the randomized rule Q is E_{z∼D}[ℓ(Q,z)] = E_{h∼Q}[L_D(h)] = L_D(Q)

Initializing...

§31.1: by linearity of expectation the generalization loss of the randomized rule Q is E_{z∼D}[ℓ(Q,z)] = E_{h∼Q}[L_D(h)] = L_D(Q) · Prove2Me