HUM AI SCI DA Damodaran Selvaraj Loss, Cost, Bias, Variance & Regularization When learning machine learning, many people get stuck not because the math is too hard, but because the ideas feel scattered.