Bayesian Inverse Problems

This course covers the use of Bayesian methods for the solution of inverse problems. Students will learn the theory and practice of Bayesian methods for parameter identification in differential equations.

Instructor: Prof. Mark Asch

Term: Autumn

Location: INRIA, Pau

Course Overview

This course provides a comprehensive introduction to Bayesian approaches for solving direct and inverse problems. Students will:

  • Learn the theoretical basis of Bayesian inference.
  • Gain practical experience with Monte-Carlo Markov-Chain (MCMC) methods.
  • Develop skills in uncertainty quantification.
  • Apply Bayesian methods to solve direct and inverse problems.

Prerequisites

  • Basic programming knowledge (preferably in Python)
  • Introductory statistics
  • Comfort with basic probability and statistics

Textbooks

  • “Bayesian Data Analysis” by Murray Gelman, et al.
  • “A Toolbox for Digital Twins” by Mark Asch

Grading

Schedule

Week Date Topic Materials
1 Introduction to inverse problems and data assimilation

Overview of the key concepts.

2 Bayesian inverse problems

Formulation and solution of stochastic inverse problems.

3 Posterior estimation methods

Techniques for performing estimation of the Bayesian posterior.