---

EECS 553 Lecture Slides
(No guarantee that they represent exactly what was covered in class.)

• l01 (intro, k-NN)
• l02 (loss, risk, Bayes classifier)
• l03 (Plug-in classifier, LDA)
• l04 (QDA, Naive Bayes)
• l05 (Regression, ERM, surrogate losses, optimization preview)
• l06 (Optimization: Lipschitz, local/global, Hessian, convex)
• l07 (Optimization methods: GD, rates, fixed point)
• l08 (Newton, SGD, logistic regression)
• l09 (Logistic regression, multiclass, imbalanced data, hyperplanes)
• l10 (Max-margin hyperplanes, constrained optimization)
• KKT / Dual function example
• l11 (KKT conditions, soft-margin hyperplane)
• l12 (Kernel methods, SVM)
• Exam1 review
• l13 (Kernel nearest centroid, Kernel ridge regression, NN)
• l14 (Backpropagation, CNN)
• l15 (Feature learning, self-supervised learning, transformer/attention)
• l16 (Self-attention, hyperparameter selection, holdout)
• l17 (hyperparameter selection, Bayesian optimization, cross validation)
• l18 (Unsupervised learning, embedding, dim. red., PCA)
• l19 (Nonlinear dimensionality reduction, t-SNE, clustering, k-means)
• Exam2 review
• l20 (GMM, latent variable models, EM)
• l21 (EM for GMM, decision trees)

---

Back to Fessler home page