The aim of the conference is to showcase the work done by the students in the Statistical Applied Mathematics at Bath (SAMBa) CDT, which encompasses research at the interface of statistics, probability, numerical analysis and applied mathematics. The conference is organised by the student cohort and will display our work to the rest of the department, mathematicians from other universities, and our industrial partners. As such, the program will be primarily devoted to talks and poster presentations from SAMBa students, complemented by invited talks from three experienced researchers from academia and industry.
If you have any questions about the conference, you're welcome to email the organisers Ben Robinson and Adwaye Rambojun.
If you would like to attend, please register by Friday 2nd June.
Registration is free but it will be very helpful for us to know numbers in advance.
Below is the provisional program for the conference. Please check back for any updates.
Tuesday 20th June | Wednesday 21st June | |
---|---|---|
9:00 -- 9:30 | Coffee | |
9:30 -- 11:00 | Session 3 - Short Student Talks | |
10:00 -- 10:30 | ||
10:30 -- 11:00 | Coffee & Discussion | |
11:00 -- 11:30 | Coffee | |
11:30 -- 12:00 | Session 1 - Uncertainty Quantification Sponsored by Schlumberger Student Talk - Owen Pembery Student Talk - Marcus Kaiser Invited Talk - Colin Morice |
Session 4 - Mathematics & Statistics in Health Sponsored by AstraZeneca Student Talk - Matt Thomas Student Talk - Robbie Peck Invited Talk - Richie Gill |
12:00 -- 12:30 | ||
12:30 -- 13:00 | ||
13:00 -- 13:30 | Lunch | Lunch |
13:30 -- 14:00 | ||
14:00 -- 14:30 | Session 2 - Networks & Applied Probability Sponsored by BT Student Talk - Anna Senkevich Student Talk - Sam Moore Invited Talk - Anja Sturm |
Session 5 - Applied Mathematical Modelling Sponsored by GKN Student Talk - Kate Powers Student Talk - Beth Boulton Close |
14:30 -- 15:00 | ||
15:00 -- 15:30 | ||
15:30 -- 16:00 | Coffee | |
16:00 -- 16:30 | Poster Session | |
16:30 -- 17:00 | ||
17:00 | Drinks Reception | |
19:00 | Dinner |
TBC
TBC
Stochastic interacting particle systems describe the time evolution of particle systems that are distributed on a discrete space (such as the integer lattice). The random changes to these systems are local, and one of the main points of interest is to understand the global and long-time behaviour that arises from a particular set of local interaction rules.
In this talk, we will discuss properties such as long-time survival and the existence of (nontrivial) invariant laws for a particular model called the cooperative branching coalescent. Here, particles perform independent random walks, only pairs of particles occupying neighbouring sites can produce new particles (cooperative branching) and particles that land on an occupied site merge with the particle present on that site (coalescence). Depending on the underlying space this system undergoes a phase transition as the branching rate is increased: For small branching rates the so called upper invariant law is trivial and the process started with finitely many particles ends up with a single particle with probability 1. Both statements are not true for high branching rates.
The talk is based on joint work with Jan Swart (UTIA Prag) and work in progress with Tibor Mach (University of Göttingen).
Macroparasites are complicated little blighters which cause infections which, due to the unconventional life cycles of parasites and the importance of variation in the severity of infection, can't be modeled like your standard bout of plague. This talk will explain why the classic SIS and SIR modelling falls apart when faced with macroparasitic infections and introduce an alternative method of modelling them, and why these models are important.
In many applications one is interested in sampling from probability distributions known up to a multiplicative constant. A standard technique to estimate such distributions are Markov Chain Monte Carlo (MCMC) methods, like Metropolis Hastings, where one generates a large number of uncorrelated samples of the distribution.
An important question is the rate at which these samples convergence to the steady state. We can think of this as the efficiency of the sampling method, which has a direct influence on the computational cost. To improve this efficiency, it is important to gain a better understanding of the involved stochastic processes.
On that account, we investigate in this talk the question why irreversible processes tend to converge faster to equilibrium than reversible ones.
Recent work in the physics literature has explored the two-species contact process as a model of staged infections. The work has a biological interpretation in terms of host-parasite invasions, for example, when a growing colony of bacteria is under threat from a developing bacteriophage infection. The problem has previously been explored from a deterministic angle, in the setting of the lattice.
In this talk we shall begin by introducing the two-species contact process in generality before moving on to suggest a novel stochastic approach to its analysis using a branching model. In particular we aim to show how a generalisation of the branching process, with a multi type offspring distribution, may be adopted to describe the evolution of the secondary wave of infection directly.
"Should I take my umbrella with me today?"
Many problems one encounters in life involve making decisions, including ones about the evidence you might want to collect to help inform further decisions. Making optimal decisions may depend on unknown parameter values, about which one may hold particular degrees of belief. To make these optimal decisions, we need to be able to formulate the problem clearly, and then apply mathematical reasoning and optimisation.
The Bayesian Decision Theory approach is based on quantifying the trade-offs between various decisions using current beliefs in probability and the costs that accompany such decisions. At the very least, the approach can be considered as making common sense deductions which are validated by rigorous mathematics by framing a problem in probabilistic terms.
After introducing the motivating examples and theory, the application to experimental design is briefly discussed, where Bayesian Decision Theory is used to help identify, and build sufficient evidence of the efficacy of, treatments that may be used to treat medical conditions.
Uncertainty Quantification (UQ) for acoustics problems is valuable for improving safety and in areas such as seismic imaging.
However, finding numerical solutions to the Helmholtz equation (the simplest model of acoustic wave propagation) is a very demanding computational task.
If we want to discover properties of sound waves moving in random media we may need to solve the Helmholtz equation thousands of times, which is computationally prohibitive, unless we can find methods which speed up our computations.
One approach to reducing the computational time needed to solve these equations is to use a preconditioner, which transforms our hard-to-solve problem into one that is easier to solve.
In this talk we will introduce current work on preconditioning for the Helmholtz equation in random media, motivated by the study of sound waves moving through random media.
In particular, we will show how, rather than needing to calculate thousands of preconditioners when we need to solve the Helmholtz equation thousands of times, in certain situations we can calculate one preconditioner, which we can then use every time we solve the Helmholtz equation, which leads to computational savings.
Nearly all modern vehicle turbochargers are formed of a centrifugal compressor and turbine. Understanding how turbochargers act under various flow regimes is important for many reasons, including improving vehicle fuel efficiency. Current modelling of turbocharger performance is predominantly via CFD, which fails to capture system dynamics near unstable flow regimes.In this talk I will describe a more analytical method for modelling turbocharger compressors. I will formulate Euler equations in this setting and use averaging techniques to reduce the system to a 1-dimensional description of the flow. Finally, I will present results for the steady-state case.
This talk will describe the asymptotic behaviour of the reinforced branching processes with fitness on the example of the preferential attachment tree of Bianconi and Barabasi. In this random graph model a popularity of a node is determined by its degree and a so-called fitness value, drawn from a specified distribution. The dynamics of these networks depend on the properties of the distribution, leading to three distinct behaviours, namely non-condensation phase, and extensive and non-extensive condensation phases. Such model provides a useful tool for understanding the growth characteristics of complex networks such as the Internet, social networks and online communities.
Calculating the burden of disease attributed to air pollution requires accurate estimation of population level exposures to pollutants. Although coverage of ground monitoring networks is increasing, these data are insufficient to independently estimate exposures globally. Information from other sources, such as satellite retrievals, chemical transport models and land use covariates must therefore be used in combination with ground monitoring data. Each of these data sources will have their own biases and uncertainties that may vary over space. Set within a Bayesian hierarchical modelling framework, the recently developed Data Integration Model for Air Quality (DIMAQ) integrates data from multiple sources and allows spatially-varying relationships between ground measurements and other factors that estimate fine particulate matter (PM2.5) concentrations. The outputs of the model are estimated exposures that can be combined with population estimates to produce population-level distributions of exposures for each country. DIMAQ was used to estimate exposures of PM2.5, together with associated measures of uncertainty, on a high-resolution grid (~11 km × 11 km) covering the entire globe for use in the 2016 WHO report ‘Ambient air pollution: A global assessment of exposure and burden of disease’, and in the 2015 and 2016 updates of the Global Burden of Disease.
For 2015, 92% of the world’s population lived in areas that exceeded the WHO 10 µg/m3 guideline. Fifty percent of the global population resided in areas with PM2.5 concentrations above the WHO Interim Target 1 (IT-1 of 35 µg/m3 ); 64% lived in areas exceeding IT-2 (25 µg/m3 ); and 81% lived in areas exceeding IT-3 (15 µg/m3 ). Nearly all (86%) of the most extreme concentrations (above 75 µg/m3 ) were experienced by populations in China, India, Pakistan, and Bangladesh.
The organisers would like to thank our industrial sponsors who helped to make this event possible. Thanks go to:
AstraZeneca, for sponsoring our session on Mathematics & Statistics in Health;
BT, for sponsoring our session on Networks & Applied Probability;
GKN, for sponsoring our session on Applied Mathematical Modelling;
Schlumberger, for sponsoring our session on Uncertainty Quantification.
This conference is also supported by the EPSRC Centre for Doctoral Training in Statistical Applied Mathematics at Bath (SAMBa), under the project EP/L015684/1.
The conference will be held in the Wolfson Lecture Theatre (4 West 1.7), in the Department of Mathematics (marked as 4W on this campus map).
The address is:
Department of Mathematics
4 West
University of Bath
Claverton Down
Bath
BA2 7AY
See the university's guide on how to find us.
If you are arriving by train, you should arrive at Bath Spa train stations. You can then take the U1, 17 or 18 bus to the university from Dorchester Street (opposite the station to the left), or walk to the university in 30-40 minutes (up quite a steep hill!).
Coming from Bristol airport, take the A4 bus to Bath, which brings you to the train station, from where you can walk or take a bus, as described above.