autumnschool

Solving Inverse Problems with Deep Learning - Autumn School

The autumn school aims to provide an introduction into solving inverse problems with deep learning methods. A background to classical inverse problems theory and reconstruction methods will be discussed and possible shortcomings. We will discuss how deep learning can be used to formulate a learned reconstruction operator to alleviate some of these shortcomings. A special emphasis is made on robustness and mathematical guarantees of the learned reconstructions.

Lecturer and topics:

Simon Arridge (UCL): Inverse Problems theory

Carola-Bibiane Schönlieb (U. Cambridge): Variational methods and learned reconstructions

Andreas Hauptmann (U. Oulu): Deep Learning basics and learned reconstruction operators

Jonas Latz (U. Manchester): Statistical Learning and Bayesian Inverse Problems

Sign-up:

Due to limited space, please send an e-mail to inversewithDL@gmail.com with CV and a short paragraph on your motivation to attend and learning goals. We will get back to you shortly after.

DEADLINE for applications: 18th of August!

Venue

The autumn school will be located at Jesus College, Jesus Ln, Cambridge CB5 8BL.

Lecture will be:

Wednesday and Thursday at the Webb Library in West Court

Friday in the Elena Hall in West Court

Schedule

We are starting wednesday at 10 am with coffee. (No early morning lecture)

We will end Friday with the afternoon coffee.

Daily schedule:

Lecture times: 9:15-10am & 10:30-12pm & 2-3:30pm

Computational session: 1-2pm & 4-5pm

Coffee breaks: 10-10:30am & 3:30-4pm

Lunch break: 12-1pm (lunch is not provided)

Social program

We will have a joint dinner at Jesus College on Thursday 21st, starting 7:30 pm.