Select the correct option[s}. Note that there might he multiple correct options. 1. [5 points) \"my is training error generally lower than validation error even on large randomly split datasets?: A. Training data are always easier than validation data. E. Models tend to t to noise present in the training set but not in the valida tion set {increasing variance}. C. Models tend to lack the abilityr to t relationships on the training dataset [increasing bias}. D. This does not happen in practice. 2. [5 points} Which of the following models should you expect to have the greatest variance? Here, i- represents a learned parameter and :r is a scalar feature: A=s s=s+sm {3-111}: n+ll+232 13- fig) = .30 + .513 + .3232 + .3333