Optimal Experimental Design and Inverse Problems
Date: 14 – 15 March 2017
Time: 9:30 – 19:00 / 9:30 – 15:30
Please note that attendance on this workshop is limited. If you or a colleague are interested in attending or would like to find out more, please contact the Turing Events team.
This two-day workshop, with attendees from academia and industry, aims to address the theory and practice in data acquisition design for reduced uncertainty in Bayesian inverse problems. Such problems are ubiquitous in earth, environmental, material and biomedical sciences. The workshop will involve short presentations and exploring specific challenges through discussion and group work.
The applications of geophysical exploration, structural integrity monitoring and material characterisation involve analytics for large sets of measurements acquired on a network of sensors. In the quest to make inferences about the surrounding media, these noisy data are subsequently processed by formulating and solving an appropriate inverse problem. This workshop aims to investigate how the statistical knowledge of a prior model can influence the data acquisition strategy, in order to maximise the expected information gain from a finite set of measurements. The workshop will involve short presentations and the exploration of specific challenges through discussion and group work.
- Lior Horesh, IBM
From Big Data to Right Data – A Hybrid First-Principles Machine-Learning Meta-Level Optimization Approach
- Alen Alexanderian, North Carolina State University
Scalable Methods for Optimal Design of Experiments for Large-Scale Bayesian Inverse Problems Governed by PDEs
- Dave C. Woods, University of Southampton
Bayesian Optimal Design for Physical Models Derived from Ordinary Differential Equations
- Paul B. Wilkinson, British Geological Survey
Adaptive Optimal Survey Design for Time-Lapse Geoelectrical Monitoring
- Jonathan Midgley, Petroleum Geo-Services
The Future of Marine CSEM – An Industry Perspective
- Johan Mattsson, Petroleum Geo-Services
Inversion of Towed Streamer EM Data for Sub Surface Resistivity
- Antonis Giannopoulos
- Tanja Tarvainen
Bayesian Approach to Quantitative Photoacoustic Tomography
- Carola-Bibiane Schöinleb
Bilevel Learning for Variational Regularisation Models
- Armin Eftekhari
Matrix Completion with Prior Information
This workshop is led by Nick Polydorides, Senior Lecturer at the University of Edinburgh and Faculty Fellow at The Alan Turing Institute.
This event is funded by the Lloyd’s Register Foundation Programme for Data-Centric Engineering.