Consider a 2-class classification problem in a 2-dimensional feature space. The training data comprise 6 sampies as shown in the following hgore (3 red triangles for the positive class and 3 green circles for the negative class). Imagine training a Thard-margin" support vector machine on the data using the defaut kerne function K(a b) = a'b. Consider what would happen if a single point were to be shifted a timy amount away from the separator found by the SVM. Which of the following points cannot be shifted away without affecting the SVM solution? (Select all correct answers.) Point 1 Point 6 Point 2 point 5 Point 4 Point 3