Machine Learning Course Calendar

COMP 388/488: Machine Learning, Fall 2015 at Loyola University Chicago

Links to: Syllabus, Grade sheet, Piazza site, Trello Board, and Bishop textbook

All future topics will always be tentative. Consult the Loyola Fall Academic Calendar for relevant registration, add/drop, and withdrawal dates.

DATE

TOPIC

EXAMS AND QUIZZES

ASSIGNMENTS AND PROJECTS

READINGS AND TUTORIALS

Aug 27 - week 1

Introduction to Machine Learning. [pdf - slides] [gDoc - lesson plan] Python environment basics [mini-course intro link]

Quiz 1: Intro to Machine Learning [sakai]

HW1: Project brainstorming and ad assignment [pdf]

Bishop 1.1 [pdf] (optional: in-class digit recognition activity [notebook])

Sep 3 -

week 2

Supervised learning, part 1: naive bayes, k-NN, linear and logistic regression [gDoc - lesson plan]

Quiz 2: supervised learning, basic [sakai]

HW2: Naive Bayes or k-NN implementation for digit recognition [gDoc]

Bishop 1.2 [pdf]. k-NN [wiki]. Naive Bayes [wiki]. Regression in sklearn 1.1.1-3 [link]

Sep 10 -

week 3

Model validation and selection [gSlides]

Quiz 3: model validation and selection [sakai]

Project discussion and group formation.

HW3: Project Proposal and HW2 extension [gDoc]

Bishop 1.3-5 [pdf], (optional: sklearn model selection docs 3.1 & 3.3 [link])

Sep 17 -

week 4

Supervised learning, part 2: SVM, random forest, neural networks [gSlides]

Quiz 4: supervised learning, advanced  [sakai]

HW4: Model selection and validation [gDoc]

see lecture slides and related section in the scikit-learn documentation for use (optional: wiki entries on random forest, neural networks, and SVMs - lead in with decision trees and perceptron)

Sep 24 - week 5

Feature selection and feature engineering [gSlides]

Quiz 5: feature engineering [sakai]

HW5: Take home Competition #1 [link] [invite code] (deadline Oct 15th)

ML mastery feature selection [html] and feature engineering [html] articles. (optional: sklearn data transform docs 4.2 and 4.3 [link])

Oct 1 -

week 6

EXAM I and

Ensemble methods: Bagging, Boosting [gSlides]

EXAM I and

Quiz 6: ensemble methods [sakai]

HW5: ongoing take home competition [link] [invite code] (deadline Oct 15)

Bagging [wiki] and Boosting [wiki]. sklearn docs on ensemble methods [html] - focus on bagging and boosting.

Oct 8 -

week 7

Dimensionality reduction: PCA, ICA [gSlides]

(Starting jointly with ChiPy: room 209 Corboy at 7pm)

Quiz 7: dimensionality reduction [sakai]

HW7: Project update and consultations [gDoc]

Intro to PCA [link]. Intro to ICA [pdf] (optional: sklearn decomposition docs [link])

Oct 15 -

week 8

Clustering: K-means and DBSCAN [gSlides]

Quiz 8: clustering [sakai]

(HW5 take home competition deadline)

HW8: Unsupervised learning assignment [gDoc]

K-means simulations and intro [link], sklearn clustering docs [link]

Oct 22 -

week 9

Semi-supervised learning: label spreading (discussion in-class)

Quiz 9: semi-supervised learning [sakai]

continued: HW7 in class, and HW8 outside of class

Semi-supervised learning [wiki] and [sklearn docs] (optional:  Intro to SSL [pdf])

Oct 29 -

week 10

Reinforcement learning [gSlides]

Quiz 10: reinforcement learning [sakai]

HW10: Tic-Tac-Toe

with reinforcement learning [gDoc] (due Nov 19)

Reinforcement learning [wiki], Flappy birds RL [link] (optional: Section 1 of [pdf] or Sutton and Barto [link])

Nov 5 -

week 11

EXAM II

In-class interview quiz [gDoc]

Nov 12 -

week 12

Bayesian networks and markov models [gSlides]

Quiz 12: Bayes nets [sakai]

HW10: code and expectations [gDoc] (due Nov 19)

Bayes Nets [pdf]

Nov 19 - week 13

Deep Learning [gSlides]

Quiz 13: Deep learning [sakai]

HW13: Deep Learning with TensorFlow (now as extra credit) [gDoc]

Deep Learning nature review [pdf]. (from the HW: the CNN TensorFlow tutorial [link])

Nov 26

Happy Thanksgiving Break!

Dec 3 - week 14

Exam III

HW 14a&b: Final project report [gDoc] and presentation [gDoc] expectations.

Please complete all online sakai quizzes and review the grade sheet.

Dec 10

Project Presentations

All assignments are due before presentations begin - NO EXCEPTIONS