The cost function in the Support Vector Machine 1 point ensures that each training example is on the decision boundary penalizes errors on query examples ensures that each example in a linearly separable training set is on the right side of the decision boundary