Syllabus (tentative)
- Introduction: what is machine learning, supervised, reinforcement and unsupervised learning
- Supervised learning/prediction: types of predictors, bias-variance tradeoff, overfitting, curse of dimensionality
- Families of predictors for regression and classification: linear and logistic regression, nearest neighbor classifiers, regression and decision trees, neural networks, boosted predictors
- Training predictors; stochastic gradient descent
- Regularization: Lasso and sparsity, Support Vector Machines and the kernel trick, overparametrization and double descent
- Learning theory
- Model selection
- Unsupervised learning problems, [semisupervised learning]
- Clustering, mixtures distributions, K-means and EM algorithms
- [Non-parametric clustering], spectral clustering
- [Markov decision processes and reinforcement learning]
Topics in [] are optional, time permitting;
new topics are highlighted in blue.
|