Bayesian Data Analysis Global South (GSU) 2021

Lecturer: Aki Vehtari

  1. Max 300 students with priority for global south and other underrepresented groups (GSU).
  2. From 4th March (first assignment deadline 12th March) to 28th May .
  3. All the material (textbook, videos, assignments, extra reading material) are freely available (see below) so you can also self-study in your own pace.
  4. The course is free (no cost) and possible to organize with help of volunteer TAs.
  5. This BDA course instance is aimed to support learning with peer support. By following the videos and doing assignments at the same time with others, you can discuss the material in assignments in the course slack, there is peer-grading platform to get feedback about your assignment solutions, and voluntary TAs help answering questions. As everything is volunteer based we can’t guarantee quick responses, but at least you will get something more than when studying only by yourself.
  6. This course is not the easiest Bayesian course available in internet, but it can be your first Bayesian course if your mathematical and programming skills are sufficient. See the prerequisites below. For easier material to start with see the end of Prerequisites section below.
  7. You will not get a formal certificate for passing the course from Aalto University.
  8. The communication happens in the course slack, please don’t email the lecturer or TAs. The slack link has been emailed to the accepted students.

All the course material is available in a git repo and via these pages are for easier navigation. All the material can be used in other courses. Text and videos licensed under CC-BY-NC 4.0. Code licensed under BSD-3.

Registration

There has been some dropouts and the registration has been re-opened here.

TA Registration

We have some volunteer TAs already, but a a few more would be great. All TAs will get a personal certificate from the lecturer Aki Vehtari if they actively participate helping students, answering questions, and possibly organize some TA sessions. We assume to have enough TAs that no-one needs to take part every week of the course and you can drop out if other obligations require so. The lecturer will support TAs.

To register as volunteer TA fill in your information (email, country, prerequisites check, a brief comment on justification why you think you can be a TA) here.

Prerequisites

  • Basic terms of probability theory
  • Some algebra and calculus
  • Basic visualisation techniques (R or Python)

This course has been designed so that there is strong emphasis in computational aspects of Bayesian data analysis and using the latest computational tools.

If you find BDA3 too difficult to start with, I recommend

Communication channels

  • The primary communication channel is the course slack. The slack link has been emailed to the accepted students.
    • Don’t ask via email or direct messages. By asking via common channels in the course chat, more eyes will see your question, it will get answered faster and it’s likely that other students benefit from the answer.
    • In the chat system, there are separate channels for each assignment and the project.
    • Channel #general can be used for any kind of general discussions and questions related to the course.
    • All important announcements will be posted to #announcements (no discussion on this channel).
    • Any kind of feedback is welcome on channel #feedback.
    • We have also channels #r, #python, and #stan for questions that are not specific to assignments or the project.
    • The lecturer and teaching assistants have names with “(staff)” or “(TA)” in the end of their names.
  • The lecturer will answer weekly the best questions about the material, either in text, recording additional video clips, or in live Q&A session (in this course instance no guarantee for weekly live session)
  • In this course instance all the TAs are volunteers from different time zones, but there may be possibility for live “TA sessions” with TAs depending on the volunteers
  • If you find errors in material, post in #feedback channel or submit an issue in github.
  • Peergrade alerts: If you are worried that you forget the deadlines, you can set peergade to send you email when assignment opens for submission, 24 hours before assignment close for submission, assignment is open for reviewing, 24 hours before an assignment closes for reviewing if you haven’t started yet, someone likes my feedback (once a day). Click your name -> User Settings to choose which alerts you want.

Assessment

Assignments (67%) and a project work with presentation (33%). Minimum of 50% of points must be obtained from both the assignments and project work. But as in this course there is no formal certificate, the assignment scores are just for your own self-evaluation. The biggest benefit from the course is the support and feedback from other students and volunteer TAs.

We use peergrade.io for providing peer feedback for the assignments and the project work. See more information at Assignments.

Schedule 2021

The course consists of 12 blocks from March to May 2021. The blocks don’t match exactly specific weeks. For example, it’s good start reading the material for the next block while making the assignment for one block. There are 9 assignments and a project work with presentation, and thus the assignments are not in one-to-one correspondence with the blocks. The schedule below lists the blocks and how they connect to the topics, book chapters and assignments.

Schedule overview

Here is an overview of the schedule. Scroll down the page to see detailed instructions for each block. Remember that blocks are overlapping so that when you are working on assignment for one block, you should start watching videos and reading text for the next block.

Block Readings Lectures Assignment Assignment due date
1. Introduction BDA3 Chapter 1 Computational probabilistic modeling,
Introduction to uncertainty and modelling,
Introduction to the course contents		 Assignment 1,
Rubric questions		 14 March
2. Basics of Bayesian inference BDA3 Chapter 1,
BDA3 Chapter 2		 Lecture 2.1,
Lecture 2.2,
Optional:
Extra explanations 2,
Summary 2.1,
Summary 2.2,
Summary 2.3		 Assignment 2,
Rubric questions		 19.03.
3. Multidimensional posterior BDA3 Chapter 3 Lecture 3 Assignment 3,
Rubric questions		 26.03.
4. Monte Carlo BDA3 Chapter 10 Lecture 4.1,
Lecture 4.2		 Assignment 4,
Rubric questions		 09.04.
5. Markov chain Monte Carlo BDA3 Chapter 11 Lecture 5.1,
Lecture 5.2		 Assignment 5,
Rubric questions		 16.04.
6. Stan, HMC, PPL BDA3 Chapter 12 + extra material on Stan Lecture 6.1,
Lecture 6.2		 Assignment 6,
Rubric questions		 23.04.
7. Hierarchical models and exchangeability BDA3 Chapter 5 Lecture 7.1,
Lecture 7.2		 Assignment 7,
Rubric questions		 30.04.
8. Model checking & cross-validation BDA3 Chapter 6, BDA3 Chapter 7, Visualization in Bayesian workflow, Practical Bayesian model evaluation using leave-one-out cross-validation and WAIC Lecture 8.1,
Lecture 8.2		 Start project work
9. Model comparison and selection BDA3 Chapter 7 (not 7.2 and 7.3),
Practical Bayesian model evaluation using leave-one-out cross-validation and WAIC		 Lecture 9.1,
Optional:
Lecture 9.2,
Lecture 9.3		 Assignment 8,
Rubric questions		 07.05.
10. Decision analysis BDA3 Chapter 9 Lecture 10.1 Assignment 9,
Rubric questions		 14.05.
11. Normal approximation, frequency properties BDA3 Chapter 4 Lecture 11.1,
Lecture 11.2
12. Extended topics Optional: BDA3 Chapter 8,
BDA3 Chapter 14-18,
BDA3 Chapter 21		 Optional:
Lecture 12.1,
Lecture 12.2		 Project work 21.05.
13. Project feedback Project feedback 28.05.

1) Course introduction, BDA 3 Ch 1, prerequisites assignment

Course practicalities, material, assignments, project work, peergrading, TA sessions, prerequisites, chat, etc.

2) BDA3 Ch 1+2, basics of Bayesian inference

BDA3 Chapters 1+2, basics of Bayesian inference, observation model, likelihood, posterior and binomial model, predictive distribution and benefit of integration, priors and prior information, and one parameter normal model.

3) BDA3 Ch 3, multidimensional posterior

Multiparameter models, joint, marginal and conditional distribution, normal model, bioassay example, grid sampling and grid evaluation. BDA3 Ch 3.

4) BDA3 Ch 10, Monte Carlo

Numerical issues, Monte Carlo, how many simulation draws are needed, how many digits to report, direct simulation, curse of dimensionality, rejection sampling, and importance sampling. BDA3 Ch 10.

5) BDA3 Ch 11, Markov chain Monte Carlo

Markov chain Monte Carlo, Gibbs sampling, Metropolis algorithm, warm-up, convergence diagnostics, R-hat, and effective sample size. BDA3 Ch 11.

6) BDA3 Ch 12 + Stan, HMC, PPL, Stan

HMC, NUTS, dynamic HMC and HMC specific convergence diagnostics, probabilistic programming and Stan. BDA3 Ch 12 + extra material

7) BDA3 Ch 5, hierarchical models

Hierarchical models and exchangeability. BDA3 Ch 5.

8) BDA3 Ch 6+7 + extra material, model checking, cross-validation

Model checking and cross-validation.

9) BDA3 Ch 7, extra material, model comparison and selection

PSIS-LOO, K-fold-CV, model comparison and selection. Extra lecture on variable selection with projection predictive variable selection.

10) BDA3 Ch 9, decision analysis

Decision analysis. BDA3 Ch 9.

11) BDA3 Ch 4 + extra material, normal approximation, frequency properties

Normal approximation (Laplace approximation), and large sample theory and counter examples. BDA3 Ch 4.

  • Read Chapter 4
  • Watch Lecture 11.1 on normal approximation (Laplace approximation) and Lecture 11.2 on large sample theory and counter examples. BDA3 Ch 4.
  • No new assignment. Work on project.
    • Review Assignment 9 done by your peers before 23:59 UTC+3 19 May, and reflect on your feedback

12) extra material + overview of BDA3 Ch 8, 14-18, 21

Frequency evaluation of Bayesian methods, hypothesis testing and variable selection. Overview of modeling data collection, BDA3 Ch 8, linear models, BDA Ch 14-18, lasso, horseshoe and Gaussian processes, BDA3 Ch 21.

  • These lectures are optional, but especially the lecture on hypothesis testing and variable selection is useful for project work.
  • Watch Lecture 12.1 on frequency evaluation, hypothesis testing and variable selection and Lecture 12.2 overview of modeling data collection, BDA3 Ch 8, linear models, BDA Ch 14-18, lasso, horseshoe and Gaussian processes, BDA3 Ch 21.
  • Work on project. TAs help with projects. Project deadline 21 May 23:59 UTC+3

13) Project evaluation

  • Project report deadline 21 May 23:59 UTC+3 (submit to peergrade).
    • Review project reports done by your peers before 26 May 23:59 UTC+3, and reflect on your feedback

R and Python

We strongly recommend using R in the course as there are more packages for Stan and statistical analysis in R. If you are already fluent in Python, but not in R, then using Python may be easier, but it can still be more useful to learn also R. Unless you are already experienced and have figured out your preferred way to work with R, we recommend

See FAQ for frequently asked questions about R problems in this course. The demo codes provide useful starting points for all the assignments.