Explores the use of deep learning for the automated creation and analysis of game metrics. Uses convolutional neural networks (CNNs) to segment and identify anything on a game screen in real-time, which is used as input to AI systems. The second part of the course analyzes the importance of the metrics. Covers surrogate models, Shannon entropy, Individual Conditional Expectation (ICE), leave-one-covariate-out (LOCO), local feature importance, partial dependency plots, tree-based feature importance, standardized coefficient importance, accumulated local effects (ALE) plots, and Shapley values. Lastly, covers building predictive models with game data using the following techniques: supervised learning, generative/discriminative learning, parametric/nonparametric learning, neural networks, unsupervised learning reinforcement learning, and adaptive control.