SPAAM student seminar series
Organisers: Hanson Bharth, Connah Johnson, Jaromir Sant and Jack Thomas
Term 1: Tuesdays 3-4pm, MS.03
Quirin Vogel | MASDOC 4th Year | Loop Percolation
Abstract: We will talk about percolation, loops and the interplay in-between the two. A number of open questions will be introduced as well as potential routes of attacking them.
Samuel Forbes | MathSys 3rd Year | Wealth Distribution Models
Abstract: I will introduce the wealth distribution, discuss empirical data for the UK and summarise some wealth distribution models.
Finally, the SPAAM poem puzzle, we state,
Just a bit of fun - for you to procrastinate!
They’ll sometimes be easy, sometimes fascinate,
But we do hope you’ll try and investigate,
For answering correctly will surely elate!
Don’t worry if you fail, you can always wait,
With a new week comes the next iterate.
And come along for the answer – don’t be late!
Ten coins are in a two-by-five grid,
Exactly four of them are to be slid,
Behold! Five straight lines of four coins now,
But the question is: can you tell me how?
Jaromir Sant | MASDOC 3rd Year
Mathematical Genetics - A Song of Wright & Fisher
Have you ever wondered why your eyes are blue,
And not some other beautiful hue?
Or perhaps you’d like to know
Why it is that your dad’s got such a prominent brow.
Or why your brother has a chin
That could easily fit an industrial bin.
Then worry not for I shall tell a story
About genetics in all it’s shining glory.
From Darwin to Fisher, from Mendel to Wright,
Their theories started to shed some light.
Then came Kimura claiming selection was dead,
And the issue has been all but put to bed.
This and some more I shall present,
To those who choose the SPAAM seminar to attend,
On Tuesday afternoon when the clock strikes three,
Just an hour before we serve coffee and tea.
Hanson Bharth | MASDOC 3rd Year | Rolls of Castamere
For this talk I shall
Introduce a mathematical model to
Simulate a cylinder rolling along an elastic
By taking a Fourier transform we
Obtain a general solution but require a
Wiener-Hopf technique to
Learn more about the solution.
“I’ve found a really nice puzzle”, said Flo,
“Oh that's great!” exclaimed Mo,
“Come on, let’s have a go!”
“C is a circle with centre O”,
“This is exciting!” shouted Mo
“I like geometry - didn’t you know?”
“Corners chosen uniformly from C,
form a random triangle - let’s call it T”
“Probability!? This I did not foresee,
And is almost definitely not for me!”
“The probability that O is in T,
That’s exactly one half - don't you agree?”
“No!” replied Mo, “on the contrary,
It’s closer to 1 over 3!”
“You are all wrong!” interrupts Lee,
"But the actual answer alludes me,
Can you solve the mystery?"
"And how about a bonus round?"
"Can the expected area of T be found?"
And as the question began to hound
From the 3 came no more sound.
Solved by: Christian Scharrer (8 days 22 hours)
Mattia Sanna | 3rd Year PhD | Computation on Galois representations
Abstract: Galois representations have a special role in algebraic number theory. They arise from different theories for example arithmetic geometry and particluar holomorphic functions called modular forms. It is quite natural then to ask whether two such Galois representations are isomorphic, and therefore provide a connection among several areas of study. One of the most famous result in this direction is the proof of Fermat's last theorem due to Andrew Wiles in which he proves that all the Galois representations attached to elliptic curves defined over comes from the ones associated to particular modular forms. We would like to extend Wile's result in the most general setting but unfortunately it is a very hard task. However, there exist some explicit methods to check whether two given Galois representations are isomorphic. In this talk we will present a brief introduction on the subject and the ideas behind the methods.
Emma Southall | 3rd Year MathSys
Early Warning Signals of Disease Elimination
A conspiracy theorist named Lee,
Frank, a risk mitigation analyst was he,
Jane, an inventor with robots she had three,
The 3 friends and 3 robots on a riverbank,
All stood still and looking blank,
”How would we cross the river?” thought Frank,
A small boat and an oar is all they had to hand,
The opposite riverbank, they wanted to land,
Luckily, the robots could operate the boat - this was well planned,
Unluckily, the boat was small and could only carry two people/robots at a time,
This line is only here so that I can rhyme,
Let’s go back to the story of maritime,
The conspiracy theorist didn’t trust the robots at all,
He was convinced that they would cause a brawl,
The risk analyst more sceptical now,
Thought "I’m sure we can sort something - but how?"
Lee insisted that no one should,
Ever be outnumbered by robots - that would be good!
The inventor wondered why her robots were so misunderstood,
But carried on anyway and quickly solved it for good.
Can you find a solution?
How many times did Jane have to cross the river?
Solved by: Christian Scharrer (158 mins)
Michael Luya | MathSys
The Mathematics behind Rotation Invariant Basis Representations for Machine Learning Models within Computational Chemistry
Abstract: How much mathematics is present in chemistry? In this presentation, Michael hopes to answer this question from a background of mathematics from the perspective of his PhD project, which is based in the design of machine learning infrastructures in computational chemistry. We will be looking at a variety of mathematical models that provide us with symmetry invariant parameters, that specifically satisfy rotation invariance, deriving them from quantum mechanical principles. Quantum mechanics itself provides us with a high level of accuracy, which we are currently interfacing with machine learning methods, which will allow us to make a variety of calculations. This talk however, will focus on the mathematics that is integral in getting us to such a stage.
Connah Johnson | MathSys
Spatial inhomogeneities in biochemical reaction networks
Abstract: Living systems may be thought of as perturbations in a chemical space. Of the set of possible chemical combinations a select few occur within biology, held within specific ranges of environments. These environments themselves being host to maelstrom of chemical interactions which may be modeled as a reaction network. Some ways in which these physical and chemical aspects combine and subsequent emergent phenomena will be the subject of this talk.
I spent a £5 note on 2 ice creams,
They were Chocolate and Caramel Supremes,
At £1.20 each, this was certainly the dream,
The number of coins in change I got was 19,
At least one 2p, 5p, 20p, 50p and £1 coins could be seen,
And exactly one 1p and one 10p I got,
How many 2p coins could be in the lot?
Nicolò Paviato | MASDOC 2nd Year | Convergence to Brownian motion in a Deterministic System
Abstract: In this talk we will discuss statistical properties of dynamical systems, i.e. an application of probability theory to a deterministic case. If a system displays a "random behaviour", then both the Central Limit Theorem and Donsker's Invariance Principle are valid for regular observables; these follow from an intrinsic martingale property of the dynamic. Finally we will provide a gist on how to improve rates of convergence for a deterministic functional CLT and to find some multidimensional estimates.
I used to play noughts and crosses,
But all I got was a number of losses,
So all the rules I decided to change,
And you might find this strange,
When players play uniformly at random,
"No skill!" you cry and cannot fathom,
But see, the skill is in choosing whether to go second or first,
Can you figure out which one is best and which one is worst?
Diogo Caetano | MASDOC 2nd Year
The Cahn-Hilliard equation on an evolving surface
Abstract: In this talk, we develop some tools to pose PDEs on evolving hypersurfaces and derive the Cahn-Hilliard equation with a logarithmic potential on a 2-dimensional moving surface. We then compare this problem with the corresponding case of a fixed domain, and conclude by discussing some results regarding existence and uniqueness of (weak) solutions.
Mohammad Noorbakhsh | MathSys
Causal network discovery from climate time series
There once was a frog called Etienne,
Who used to sit on a riverbank in the Glen,
Lilly pads on water there was one shy off N,
Formed a straight line just in front of Etienne,
He chooses a number uniformly from 1 to N,
And jumps to this numbered lillypad then,
But if he chooses N, he crosses the river in the Glen.
And if he lands on a lillypad, the process repeats again,
Until on the other side of the riverbank is our dear Etienne
Given he jumps forwards on every jump, what is the expected number of jumps for Etienne to cross the river?
Letizia Angeli | MASDOC | Interacting Particle Systems Approximations of Feynman-Kac Formulae in continuous time
Abstract: In this talk, I will present a class of numerical algorithms - based on the evolution of interacting particle systems - for estimating large deviation quantities of additive path functionals of stochastic processes. Adapting already established results from the literature of particle filters and sequential Monte Carlo methods, we can study the convergence of the algorithms and provide a rigorous framework to evaluate and improve their efficiency.
Osian Shelley | MASDOC | Transactions costs, boon or bust?
Abstract: The financial crisis of 2008 has helped further the debate on the desirability of transaction taxes in financial markets. Indeed, recurring questions and themes relating to such taxation often gain relevance after periods of economic instability. Unfortunately, arguments for and against these taxes are often based on ad hoc heuristics. This is due in part to a lack of general equilibrium models, from which one can gather quantitative and qualitative guidance.
In the talk I aim to show the effects of a quadratic tax rate levied against two agents with heterogeneous risk aversions in a continuous-time, risk-sharing equilibrium model. In particular, I’ll talk about the utility loss incurred by the agents upon the introduction of the tax and whether this tax can benefit the agents upon redistribution.
Yani Pehova | Mathematics PhD
Quasirandomness in large permutations via the flag algebra method
Abstract: A sequence (Pi) of permutations is said to be quasirandom if the pattern density of every 4-point permutation in (Pi) converges to 1/24. We show that there is a set S of eight 4-point permutations such that the sum of the pattern densities of the permutations from S in (Pi) converges to |S|/24 if and only if the sequence is quasirandom. In this talk I will go over the proof methods we use to show that S has the above property. In particular, I will give an outline of the philosophy of the flag algebra method, and some details on how to apply it.
This is joint work with Timothy Chan, Jon Noel, Maryam Sharifzadeh, Daniel Kral and Jan Volec.
Phil Herbert | MASDOC | The membrane mediated force on point attachments with application to a near spherical biomembrane
Abstract: We consider a hybrid model of a biomembrane with attached particles. The membrane is represented by a near spherical continuous surface, attached proteins are described by rigid bodies which are free to move tangentially and rotate in the axis normal to a reference point. As the standard energy for a membrane is highly non-linear, we consider a quadratic energy which may be shown to be an approximation of the Canham-Helfrich energy with a volume constraint and the deformations due to the attached proteins are imposed by point Dirichlet constraints. We show differentiability of the membrane energy with respect to parameterisation of these embedded particles and provide illustrative numerical examples.
Dom Brockington | MASDOC | Sticky Flows and KPZ Universality
Abstract: We shall introduce stochastic flows, and stochastic flows of kernels, before moving to the special case of sticky flows and their fluctuations. On the large deviation scale these fluctuations turn out to be Tracey-Widom GUE distributed, thus the sticky flows lie in the KPZ universality class.
Term 2: Tuesdays 2-3pm, B3.01
We're happy to announce and formally invite everyone to the SPAAM seminar series, a new forum where Maths & Stats PhD students as well as early researchers can present and discuss their work. The series is being organised by the SIAM Student Chapter here at Warwick.
Tea & Coffee? Yes – afterwards in the Maths common room
The seminar series aims to bring together CDT streams as well as straight PhD students across Mathematics and Statistics.
If you're interested in giving a talk, please let us know!