CSC 480/580: Principles of Machine Learning - Spring 2024

Tentative schedule

Please refer to this course’s Fall 2022 version for a preview of the lecture slides.

Date	Topics	Notes	Readings	Homework
Jan 11	Introduction, motivation, course mechanics	slides	CIML Chap. 1	HW0
Jan 16	Supervised learning paradigm; decision trees	slides	CIML Chap. 2
Jan 18	Finish decision trees; Bayes classifier and its error		CIML Chap. 3	HW0 due
Jan 23	Overfitting: detection and mitigation; Hyperparameter selection	slides
Jan 25	Geometry & Nearest neighbors	slides	CIML Chap. 3	HW1
Jan 30	K-means clustering; Probability review	slides
Feb 1	HW0 review; Begin linear models		CIML Chap. 4
Feb 6	Linear models; the Perceptron algorithm	slides	CIML Chap. 5
Feb 8	Practical considerations I: features in supervised learning
Feb 13	Practical considerations II: performance measure beyond error rates
Feb 15	Practical considerations III: confidence intervals, hypothesis testing, debugging ML algorithms, bias-variance tradeoff	slides		HW1 due
Feb 20	Linear models for regression I: least squares, maximum likelihood		CIML Chap. 7
Feb 22	Linear models for regression II: regularization	slides		HW2
Feb 27	Linear models for classification: logistic regression
Feb 29	Linear models for classification: SVM; midterm review	slides
Mar 5	Spring Recess
Mar 7	Spring Recess
Mar 12	Midterm exam
Mar 14	Nonlinear models: basis functions; begin kernels	slides	CIML Chap. 11
Mar 19	Finish kernels
Mar 21	Unsupervised learning basics: clustering		CIML Chap. 15	Project proposal due Mar 22
Mar 26	Unsupervised learning: PCA	slides
Mar 28	Finish PCA; Begin Probabilistic ML			HW3 Data: three.txt eight.txt
Apr 2	Probabilistic ML: Bayes nets; Basic examples	slides	CIML Chap. 9
Apr 4	Probabilistic ML: Naive Bayes		CIML Chap. 16
Apr 9	Finish Naive Bayes; EM; Gaussian Mixture Models	slides
Apr 11	Finish EM for Gaussian Mixture Models; begin Neural Networks	slides	Stanford CS231n note: optimization
Apr 16	Neural networks: backpropagations; consideration in architecture design		Stanford CS231n note: backprop
Apr 18	Neural networks: batch normalization; Begin convolutional neural networks		Stanford CS231n note: convnets	HW4
Apr 23	Finish convolutional neural networks; Autoencoders		NanoGPT tutorial by Andrej Karpathy
Apr 25	Reinforcement learning: MDP; policy evaluation		Deep RL: Pong from pixels by Andrej Karpathy
Apr 30	Project feedback session; Final review	slides
May 7	Final exam at 8:00 – 10:00am