Machine Learning for Structured Data

10-418 + 10-618, Fall 2019
School of Computer Science
Carnegie Mellon University


Important Notes

This schedule is tentative and subject to change. Please check back often.

Lecture Videos

  • You can view all the videos (both past and scheduled live streams) here: https://www.youtube.com/channel/UCF4lXfWfgTvslz6yNGVkQTg
  • We will include links to individual videos/live streams on the schedule below.
  • See Piazza note "Welcome to the Live Stream!" for additional details.

Tentative Schedule

Date Lecture Readings Announcements

Search-based Structured Prediction

Mon, 26-Aug Lecture 1 : Course Introduction
[Slides] [Whiteboard] [Video]

Wed, 28-Aug Lecture 2 : Reducing Multiclass to Binary Classification
[Slides] [Whiteboard] [Video]

Fri, 30-Aug (No Recitation)

Mon, 2-Sep (No Class: Labor Day)

Wed, 4-Sep Lecture 3 : Structured Prediction as Search
[Slides] [Whiteboard] [Video]

Fri, 6-Sep Recitation: PyTorch

Mon, 9-Sep Lecture 4 : Learning to Search / Recurrent neural networks (RNNs)
[Slides] [Whiteboard] [Video]

Wed, 11-Sep Lecture 5 : Sequence-to-sequence Models
[Slides] [Whiteboard] [Video]

HW1 out (Thu)

Fri, 13-Sep Recitation: HW1

Graphical Models: Representation

Mon, 16-Sep Lecture 6 : Locally Normalized Models: Bayesian Networks
[Slides] [Whiteboard] [Video]

Wed, 18-Sep Lecture 7 : Globally Normalized Models: Markov Random Fields & Conditional Random Fields
[Slides] [Whiteboard] [Video]

Fri, 20-Sep (No Recitation)

Mon, 23-Sep Lecture 8 : Factor Graphs / Exact Marginal Inference: Variable Elimination
[Slides] [Whiteboard] [Video]

Graphical Models: Exact Inference and Learning

Wed, 25-Sep Lecture 9 : Exact Marginal/MAP Inference: Belief Propagation
[Slides] [Whiteboard] [Video]

HW1 due (Thu)

Fri, 27-Sep (No Recitation)

HW2 out (Sat)

Mon, 30-Sep Lecture 10 : Learning fully observable MRFs and CRFs
[Slides] [Whiteboard] [Video]

Wed, 2-Oct Lecture 11 : Neural Potential Functions
[Slides] [Whiteboard] [Video]

Fri, 4-Oct Recitation: HW2

Mon, 7-Oct Lecture 12 : MAP Inference: Mixed Integer Linear Programming (Part I)
[Slides] [Whiteboard] [Video]

Wed, 9-Oct Lecture 13 : MAP Inference: Mixed Integer Linear Programming (Part II)
[Slides] [Whiteboard] [Video]

Fri, 11-Oct Recitation: Midterm Exam Review

HW2 due (Sat)

Learning for Structured Prediction

Mon, 14-Oct Lecture 14 : Midterm Exam Review / Structured Perceptron
[Slides] [Whiteboard] [Video]

Wed, 16-Oct Lecture 15 : Structured SVM
[Slides] [Whiteboard] [Video]

Thu, 17-Oct Midterm Exam (evening exam -- details will be announced on Piazza)

Fri, 18-Oct (No class: Mid-semester break)

Approximate Inference: MCMC

Mon, 21-Oct Lecture 16 : Convolutional neural networks (CNNs)
[Slides] [Whiteboard] [Video]
  • Convolutional Networks. Ian Goodfellow, Yoshua Bengio, & Aaron Courville (2016). Deep Learning, Chapter 9.1 - 9.3.

Wed, 23-Oct Lecture 17 : Monte Carlo Methods
[Slides] [Whiteboard] [Video]
  • Monte Carlo Methods. Li (2003). Information Theory, Inference, and Learning Algorithms, Chapter 29 (Section 29.1-29.3).

Project team due

Thu, 24-Oct Recitation: HW3 (rescheduled to Mon, Oct-28 6:30pm-7:30pm in GHC 4401)

HW3 out

Fri, 25-Oct (No class: Day for community engagement)

Mon, 28-Oct Lecture 18 : Markov Chain Monte Carlo: Gibbs Sampling & Metropolis-Hastings (Recitation: HW3 at 6:30pm-7:30pm in GHC 4401)
[Slides] [Whiteboard] [Video]
  • Monte Carlo Methods. Li (2003). Information Theory, Inference, and Learning Algorithms, Chapter 29 (Section 29.4 - 29.5).

Wed, 30-Oct Lecture 19 : Markov Chains
[Slides] [Whiteboard] [Video]
  • Monte Carlo Methods. Li (2003). Information Theory, Inference, and Learning Algorithms, Chapter 29 (Section 29.6 - 29.10).

Fri, 1-Nov 10-618 Project Team Office Hours

Approximate Inference: Variational Methods

Mon, 4-Nov Lecture 20 : Bayesian Inference for Parameter Estimation
[Slides] [Whiteboard] [Video]

Wed, 6-Nov Lecture 21 : Topic Modeling
[Slides] [Whiteboard] [Video]

HW4 out

HW3 due

Fri, 8-Nov Recitation: HW4

Project team due (Thu, Nov-07)

Mon, 11-Nov Lecture 22 : Mean Field Variational Inference (Part I)
[Slides] [Whiteboard] [Video]

Project proposal due

Wed, 13-Nov Lecture 23 : Mean Field Variational Inference (Part II)
[Slides] [Whiteboard] [Video]

Fri, 15-Nov 10-618 Project Team Office Hours

Mon, 18-Nov Lecture 24 : Coordinate Ascent Variational Inference
[Slides] [Whiteboard] [Video]

HW4 due

Advanced Topics

Wed, 20-Nov Lecture 25 : Learning partially observable graphical models / Variational EM
[Slides] [Whiteboard] [Video]

HW5 out

Fri, 22-Nov Project midway poster session (Session I) [1:00pm - 3:00pm in GHC 6115]

Project midway poster due (Thu, Nov-21)

Mon, 25-Nov Lecture 26 : Bayesian Nonparametrics / Project midway poster session (Session II) [7:30pm - 9:30pm in GHC 6115]
[Whiteboard] [Video]

Wed, 27-Nov (No class: Thanksgiving break)

Fri, 29-Nov (No class: Thanksgiving break)

Mon, 2-Dec Recitation: Final Exam Review

HW5 due

Wed, 4-Dec Lecture 27 : Variational Autoencoders / Restricted Boltzman Machines
[Whiteboard] [Video]

Thu, 5-Dec Final Exam (evening exam) [6:30pm - 9:00pm in DH A302]

Fri, 6-Dec (No Recitation)

Wed, 11-Dec Project final poster session (time/location TBD)

Project final poster due (Tue, Dec-10)