STAT 339 (Spring 2020): Tentative Schedule of Topics

.

The following is a *tentative* schedule of what we will cover in class when. Check this space regularly for updates.

**Last update: 06-18-21**

Key to reading materials:

- FCML: First Course in Machine Learning (Rogers and Girolami)
- BRML:
*Bayesian Reasoning and Machine Learning*(Barber) - MLPP:
*Machine Learning: A Probabilistic Perspective*(Murphy)

Date | Topics | Readings | Assignments |
---|---|---|---|

M 2/03 | What is (Probabilistic) Machine Learning? | MLPP 1.1-1.3, BMRL 13.1 | – |

[PART I: Machine Learning Fundamentals | (~ 2 weeks)] | ||

W 2/05 | Nearest Neighbor Classification | FCML 5.3.1, BRML 14.1-14.2 | Background Survey Due |

Evaluating a Classifier | |||

F 2/07 | Cross-Validation | FCML 1.5 | – |

M 2/10 | Linear Regression: One Scalar Predictor | FCML 1.1-1.4 | – |

W 2/12 | Linear Regression: Vector Predictor | HW0: Optional Warmup due | |

F 2/14 | Linear Regression: Vector Predictor (cont’d) | – | |

M 2/17 | Linear Regression: Polynomials as Linear Models | – | |

W 2/19 | Linear Regression: Regularization | – | HW1a: Classification due |

[PART II: Probability Fundamentals | (~ 2 weeks)] | ||

F 2/21 | Probability: Probability Spaces | FCML 2.1 | – |

M 2/24 | Probability: Discrete Random Variables | FCML 2.2, BRML 1.1, 8.1-8.3 | – |

W 2/26 | Probability: Discrete Random Variables (cont’d) | FCML 2.2, BRML 1.1, 8.1-8.3 | HW1b: Regression due |

F 2/28 | Probability: Joint and Conditional Probability | FCML 2.3-2.6, BRML 1.1, 8.1-8.3 | – |

M 3/02 | Probability: Continuous Random Variables | FCML 2.3-2.6, BRML 1.1, 8.1-8.3 | – |

W 3/04 | Probability: Expectation and Variance | FCML 2.3-2.6, BRML 1.1, 8.1-8.3 | |

[PART III: Inference Fundamentals (~ 3 weeks)] | – | – | |

F 3/06 | Maximum Likelihood Estimation | FCML 2.7-2.8, BRML 8.6, 8.8.1 | – |

M 3/09 | Bias/Variance Decomposition | FCML 2.9-2.11 | – |

W 3/11 | Bias and Variance in Regression | FCML 2.9-2.11 | HW2a: Probability Basics due |

HW2b: Maximum Likelihood Estimation due | |||

F 3/13 | (Zoom / Logistics Discussion) | – | – |

3/16-3/27 | EXTENDED SPRING BREAK | ||

M 3/30 | Bayesian Inference: Fundamentals | FCML 3.1-3.3, BRML 1.2-1.4 | – |

W 4/01 | Bayesian Inference: Conjugate Priors | – | |

F 4/03 | Bayesian Inference: Exponential Families | – | |

M 4/06 | Belief Networks and Conditional Independence | FCML 3.6, BRML 3.1-3.3 | – |

W 4/08 | Bayesian Inference: Model Averaging and Occam’s Razor | HW3: Bayesian Inference and Belief Networks due | |

[PART IV: Bayesian Supervised Learning | (~ 2 weeks)] | ||

F 4/10 | Naive Bayes Classification: Formulation and Max Likelihood | FCML 5.1-5.2.1.6, BRML 10.1-10.2 | – |

M 4/13 | Bayesian Naive Bayes I: Posterior Inference | FCML 5.2.1.7, BRML 10.3 | “Midterm” Distributed |

W 4/15 | Bayesian Naive Bayes II: Prediction and Classification | – | – |

F 4/17 | Bayesian Regression I: Posterior Inference | FCML 3.8, BRML 12.1-12.4 | – |

M 4/20 | Bayesian Regression: Prediction and Model Selection | FCML 3.9, BRML 12.1-12.4 | “Midterm” Due |

[PART V: Unsupervised Learning and Latent Variable Models | (~2 weeks)] | ||

W 4/22 | Clustering: K-Means | FCML Ch. 6.1-6.2 | – |

F 4/24 | Clustering: Mixture Models | FCML 6.3.1-6.3.2, BRML 20.1 | – |

M 4/27 | Clustering: The EM Algorithm | FCML 6.3.3-6.3.8, BRML 20.2-20.3 | [HW4: Bayesian Classification and Regression due |

W 4/29 | Clustering: Gibbs Sampling | FCML 6.3.9, 9.1-9.2, 10.1 | – |

F 5/01 | Hidden Markov Models: Overview | BRML 23.1-23.3, 23.5 | – |

M 5/04 | HMMs: Prediction and Inference | – | |

W 5/06 | HMMs: Learning with EM | – | |

F 5/08 | HMMs: Bayesian Learning | – | |

W 5/13 9pm | HW5: Clustering due | – |