As part of the UDRC phase III, a first themed meeting on Scalable Signal Processing with Bayesian Graphical Models will be held in Edinburgh on Wednesday, 20th February. This event is intended for academic researchers, industrial partners and Dstl staff to learn about and discuss current trends in scalable Bayesian methods.
 
The program will consists of a series of talks covering Bayesian sensor fusion, Bayesian networks, approximate message passing and advanced Monte Carlo sampling techniques, followed by a networking session. We will also present a Dstl Challenge competition.

Download Agenda

Agenda

9:00 – 9:30 Refreshments
 
9:30 – 9:40 Introductions and welcome, Yoann Altmann, Heriot-Watt University
 
9:40 – 10:20 Plenary talk: Convergence Rates for a Class of Monte Carlo Estimators Based on Stein's Method, Mark Girolami, Imperial College London - Abstract
 
10:20 – 10:50 Sensor fusion, David Cormack, Heriot-Watt University; James Hopgood, University of Edinburgh
 
10:50 – 11:20 Advanced MCMC, Kostas Zygalakis, Heriot-Watt University

11:20 – 11:50 Refreshments
 
11:50 – 12:20 Approximate message passing and expectation propagation, Yoann Altmann, Heriot-Watt University
 
12:20 – 12:50 Scalable opportunistic calibration of fusion networks, Murat Uney, CMRE
 
12:50 – 13:50 Lunch and posters
 
13:50 – 14:20 Scaling signal processing for industry needs, David Greig, Leonardo
 
14:20 – 14:40 Dstl challenges, Jordi Barr, Dstl; Gary Heald, Dstl
 
14:40 – 15:00 Refreshments
 
15:00 – 15:45 Panel discussion and questions
 
16:00 Close

For more information contact janet.forbes@ed.ac.uk