Introduction to Bayesian Modeling

Description

Learn key concepts and practices of Bayesian modeling. Explore models with conjugate priors, illustrating with the Beta-Binomial and Gamma-Poisson models, Markov Chain Monte Carlo (MCMC), variational inference (VI) methods for Bayesian inference, and semi-conjugate models, including Bayesian linear regression, Bayesian mixture models, and Bayesian hidden Markov models. Also, dig deeper into models used when there are not conjugate priors, such as Bayesian logistic regression, Bayesian multiclass regression, a racially polarized voting model, and Bayesian deep learning.

Students completing the course will be able to:

  • Choose a Bayesian model appropriate for modeling a particular real-world situation.
  • Fit a Bayesian model to real-world data using Markov Chain Monte Carlo (MCMC) or Variational Inference methods where appropriate. 
  • Use Bayesian models to predict future behavior of a modeled system.
  • Choose between options for Bayesian models for describing a real-world situation.
  • Evaluate the accuracy and appropriateness of a Bayesian model in predicting the behavior of a real-world system.

Instructors

Dr. Michael Wojnowicz

Prerequisites

Basic Enrollment Recommendations: Students should be comfortable with calculus, linear algebra (matrix multiplications, determinants, and traces), and introductory probability (e.g., expectation, conditional probability, and commonly used statistical distributions, including Gaussian and Poisson). Students should be knowledgeable in Python.

This course will be held:

Dates: January 3-13, 2023
Format: online, with synchronous daily sessions from 12-1:30pm ET
Cost: $1697 (limited funding available for PhD candidates, please reach out to courses@tufts.edu for information)

How to enroll:

Tufts Degree Students (graduate and PhD)

Please follow the instructions below to enroll in a winter term for-credit course.

  1. Email Course Updates at course_updates@tufts.edu and request to be term activated for the Winter Session. You will receive an email back within two business days confirming your activation.
  2. Once you have been activated, log in to SIS and enroll in a winter term course.

Non-Degree Students (visiting students and working professionals)

Register via University College’s website by following the link

UC DIS 202 Matrix Methods for Machine Learning

UC DIS 211 Intro to Bayesian Modeling