Chapter 16 treats the topic of machine learning in EMC simulation. The application of machine learning is a fast-growing field, with its full potential not yet unfolded. In this chapter, we apply one concept of machine learning, namely, wideband Gaussian process regression (GPR), to generate a surrogate model of the circuit model of an automotive traction inverter. With the surrogate, we perform a multi-objective optimization (MOO) with objectives chosen to minimize the electromagnetic interference (EMI) filter of the inverter under consideration of performance and mechanical design variables. We treat in total nine design dimensions and investigate the Pareto fronts obtained from the MOO. We also introduce the scatter plot matrix and parallel coordinate plots to examine optimal solutions. In the chapter, an introduction to GPR and MOO is given. The algorithms for wideband GPR is derived at the end of the chapter.