Introduction to Machine Learning

Spring 2018 — CMSC 678

Announcements
Who, What, When, and Where

Check out the syllabus for all this information, including policies on academic honesty, accomodations, and late assignments.

Meeting Times
Performing Arts and Humanities, 132
Monday & Wednesday, 2:30pm - 3:45pm
Instructor
Frank Ferraro
ferraro [at] umbc [dot] edu
ITE 358
Monday 3:45 - 4:30
Tuesday 11:00 - 11:30
by appointment
TA
Vamshi Nagabandi
nvamshi1 [at] umbc [dot] edu
ITE 349F
Wednesday 1:00 - 2:00
Thursday 2:30 - 3:30
Topics
The topics covered will include, but are not limited to:
  • perceptrons
  • regression (linear, logistic, and non-linear)
  • spectral, clustering, and dimensionality reduction techniques
  • support vector machines and kernel methods
  • neural networks, including deep learning, recurrent neural networks, and convolutional neural networks
  • Bayesian networks and probabilistic graphical models
  • clustering
  • evaluation methodologies and experiment design.
Goals
After taking this course, you will
  • be introduced to some of the core problems and solutions of ML;
  • learn different ways that success and progress can be measured in ML;
  • be exposed to how these problems relate to those in statistics, artificial intelligence, and specialized areas of ML (such as natural language processing and computer vision);
  • have experience implementing a number of ML programs;
  • read and analyze research papers;
  • practice your (written) communication skills.
Schedule

The following schedule of topics is subject to change.

Legend:

Date Topic Suggested Reading Assignment Out Assignment Due
Monday, 1/29 Introduction: what is ML?
  • ESL Ch 1
  • ESL Ch 2
 A1: Math & Programming Review —
Wednesday, 1/31 Linear models, loss functions, and decision theory
  • CIML Ch 7
  • UML Ch 2
  • UML Ch 14
  • ITILA Ch 36
 — —
Monday, 2/5 Linear regression and perceptrons
  • ESL Ch 3
  • CIML Ch 7.6
  • UML Ch 9.2
 — —
Wednesday, 2/7 Course Project A1
Monday, 2/12 Logistic regression/maxent models
  • ESL Ch 4.4
  • UML Ch 9.3
  • CIML Ch 9.5-9.7
  • ITILA Ch 39, 41.1-41.3
 A2: Multiclass Classification —
Wednesday, 2/14 Neural networks, backpropagation, & autodifferentiation
  • ESL Ch 11
  • CIML Ch 10
  • UML Ch 20
  • ITILA Ch 38-39
 — —
Monday, 2/19 — —
Wednesday 2/21 Recurrent & convolutional neural networks — —
Monday, 2/26 Guest Lecture: Dr. Cynthia Matuszek — — —
Wednesday 2/28 Recurrent neural networks/catch up — —
Monday, 3/5 Dimensionality Reduction: Linear Discriminant Analysis & Principal Component Analysis
  • ESL Ch 4.3
  • ESL Ch 14.5-14.10
  • CIML Ch 15.2
  • UML Ch 23, 24.3
 — —
Wednesday, 3/7 Prototype vs. exemplar learning: k-means and k-nearest neighbor
  • ESL Ch 13, 14.3
  • CIML Ch 15.1
  • UML Ch 22
  • ITILA Ch 20
 — —
Friday, 3/9 — — — A2
Monday, 3/12 Kernel methods & Support vector machines
  • ESL Ch 6 (Kernel)
  • UML Ch 16 (Kernel)
  • ESL Ch 12 (SVM)
  • UML Ch 15 (SVM)
  • CIML Ch 11 (Kernel + SVM), 7.7 (SVM)
 — Project Proposal
Wednesday, 3/14 A3: Clustering —
Monday, 3/26 Exam 1 (in-class) — —
Wednesday, 3/28 Expectation Maximization
  • ESL Ch 8.5
  • UML Ch 24.0-24.1 (Maximum likelihood)
  • UML Ch 24.4 (EM)
  • CIML Ch 16
  • ITILA Ch 20
 — —
Monday, 4/2 — —
Wednesday, 4/4 Directed Probabilistic Graphical Models (PGMs)
  • ITILA Ch 16 (Message Passing, e.g., for forward-backward)
  • CIML Ch 9
 — —
Monday, 4/9 — —
Wednesday, 4/11 Undirected PGMs
  • ITILA Ch 25 and 26 (~16 pages)
  • ESL Ch 17 (excluding 17.3.2, 17.4.2-end; ~17 pages)
 A4: Probabilistic Graphical Models A3
Monday, 4/16 — Project Update
Wednesday, 4/18 PGM Inference: Variational Methods
  • ITILA Ch 33 (~15 pages)
  • ESL Ch 8.5
 — —
Monday, 4/23 — —
Wednesday, 4/25 PGM Inference: Sampling
  • ITILA Ch 29.0-29.6 (~17 pages)
  • ESL Ch 8.6
 — —
Monday, 4/30 Structured Prediction
  • CIML Ch 17
  • ESL Ch 6.4
 — —
Wednesday, 5/2 — —
Monday, 5/7 catch up — —
Wednesday, 5/9 Ensemble Methods
  • CIML Ch 13
 — —
Monday, 5/14 Reinforcement Learning — A4
Friday, 5/18 Exam 2 — —
Wednesday, 5/23 — — Final Project