Lectures for COMP-598: Topics in Computer Science: Applied Machine Learning

Fall 2014


Some lecture notes will be linked to this web page, in PDF format. The reader for PDF files is available free from Adobe for UNIX, Apple Macintosh, and Windows.

Schedule

Lec.
Date
Topic
Lecture Material
Homeworks and Readings
1
Sep. 9

Introduction to machine learning.
Read this paper.
Review basic notions of algebra and probabilities.
Read Ch.1-2 of Bishop and Ch.1 of Hastie et al.
2
Sep.11

Linear regression.
Suggested readings:
Ch.2 (Sec.2.1-2.4, 2.9) of Hastie et al.
Ch.3 of Bishop (Sec.3.1-3.2).
3
Sep. 16

Linear regression.
Mini-project #1 available. Suggested readings:
Ch.3 (Sec.3.1-3.4, 3.9) of Hastie et al.
Ch.3 of Bishop (Sec.3.1-3.2).
4
Sep. 18

Linear classification.
Suggested readings:
Ch.4 of Hastie et al.
Ch.4 of Bishop (Sec.4.1-4.3).
A paper by Ng & Jordan (NIPS, 2001).
5
Sep. 23

Naive Bayes.
Suggested readings:
Sec. 6.6.3 of Hastie et al.
Ch.13 (Sec.13.1-13.4) of the book Introduction to Information Retrieval.
6
Sep. 25

Performance analysis and error estimation.
Suggested readings:
Ch.7 of Hastie et al.
Wagstaff (2012) paper
Dworkin et al. (2015) paper
Blum & Hardt (2015) paper (more theoretical)
7
Sep. 30

Practical session with python and scikit-learn.
Mini-project #1 due.
8
Oct. 2

Decision trees
Mini-project #2 available. Suggested readings:
Sec.14.4 of Bishop.
Sec.9.2 of Hastie et al.
9
Oct. 7

Instance-based learning
Suggested readings:
Sec.2.5 of Bishop.
Sec.13.3 of Hastie et al.
10
Oct. 9

Ensemble methods
Suggested readings:
Sec.8.7, Ch.10 of Hastie et al.
Ch.14 of Bishop
11
Oct. 14

Ensemble learning (cont'd)
12
Oct. 16

Support vector machines
Suggested readings:
Sec.7.1 of Bishop.
Ch.12 (Sec.12.1-12.4) of Hastie et al.
For background on convex optimization: see this book by S. Boyd and L. Vandenberghe
13
Oct. 21

Support vector machines (cont'd)
Mini-project #2 due.
14
Oct. 23

Neural networks
Mini-project #3 available. Suggested readings
Ch.11 of Hastie et al.
Sec.5.1-5.3 of Bishop
15
Oct. 28

Neural networks (cont'd)
16
Oct. 30

Feature construction and selection
17
Nov. 4

Deep learning
18
Nov. 6

Problem solving session
19
Nov. 11

Unsupervised learning
Mini-project #3 due.
20
Nov. 13

Online / streaming data
Final project available.
21
Nov. 18

Semi-supervised learning
22
Nov. 20

Parallelization for large-scale ML
23
Nov. 25

Midterm (tentative date).
24
Nov. 27

Missing data.
25
Dec. 2

Final project presentation session
26
Dec. 4

Wrap-up
Final project report due.