Streaming Analytics Experts

LBDN’s high-velocity analytics experts are Bayesians with backgrounds in defence, epidemiological and economic applications and in improving the computational efficiency of the underpinning techniques: 

Prof Simon Maskell (SM) is adept at the Bayesian processing of high-variety large-scale streaming data, specifically for data-fusion and tracking applications, eg in defence and security where veracity is key. With over 7800 citations, SM’s tutorial 1 is the most cited paper on particle filters. SM also wrote Wiley’s CS Encyclopaedia definition of tracking and, in 2010, was general chair for Fusion 2010, the largest international data fusion conference (which hasn’t otherwise been in the UK). SM is associate editor for IEEE Signal Processing Letters and IEEE Transactions on Aerospace and Electronic Systems. He was previously a technical lead at QinetiQ, leading commercial (streaming and batch) data-analytics projects with a focus on achieving breakthroughs in performance (eg using particle filters similar to developed through intergovernmental collaboration, he delivered a fourfold improvement in sensitivity for a specific military aircraft). SM is director of the LBDN and co-organising a forthcoming public-engagement event, “Big Data or Big Brother”.

Prof Brendan McCabe (BM) is expert in modelling data such as (but not exclusively) the multivariate time-series found in Economics/Finance. The problems considered range from the Bayesian analysis of dependence among rare events to very large data sets related to financial trading (where the quantity of variables and data make the models intractable). 

Prof Ivan Au (IA) invented subset simulation. With 580 citations, and a recently published book, it represents a break-through technique to dramatically improve importance sampling (IS)’s efficiency, eg as used by particle filters. IA pioneered MCMC’s use in structural system identification, developed efficient IS strategies and helped understand how IS combats the curse of dimensionality.