Department of Mathematics

CoSy Lunch Seminars 2017 Autumn

29 August
Speaker: Alicia Fornes, Computer Vision Center, Universitat Autonoma de Barcelona
Title: Semantic recognition in historical handwritten documents
Place: Å11167
Time: 12:15 -- 13:00
Document Image Analysis and Recognition (DIAR) is the pattern recognition research field devoted to the analysis, recognition and interpretation of images of documents.

In the last decades, DIAR has become a cornerstone technology for the preservation of cultural heritage. Concretely, the extraction of relevant information from historical handwritten document collections is one of the key steps in order to make these manuscripts available for access and searches. In this context, instead of a pure transcription, the objective is to move towards document understanding (including named entity recognition and semantic categorization). A typical example can be demographic documents, where the extraction of named entities (e.g. family names, places, dates, etc.) and its storage in structured databases allows to envision innovative services based in genealogical, social or demographic searches.

This talk will present several context-aware and deep learning-based methods (e.g. BLSTMs, CNNs) for extracting information in handwritten documents.

12 September
Speaker: Raazesh Sainudiin, Department of Mathematics, Uppsala University
Title: Ancestries of a Recombining Diploid Population: towards a theoretical union of behavioural ecology and population genetics
Place: Å4004
Time: 12:15 -- 13:00
First, I will share behavioural ecological insights in a monogamous, patri-local, territorial, bi-parental system from a field work on a remote South Pacific island. Having appreciated the complexities in the field, we will start from first principles and build towards topologically explicit (karyotic, sub-karyotic, cytosolic and zygotic) notions of inheritance in the simplest mathematical models of coalescent with recombination that is embedded within behaviourally specifiable diploid population pedigree processes.

I will convey by pictures the ideas in a recent theoretical unification of previously disparate population genetic processes that underpin modern statistical methods in population genomics: (1) Kingman's coalescent trees, (2) Chang's population pedigrees and (3) Griffiths' and Hudson's ancestral recombination graphs.  Using a single parameter in the unit interval we build a family of combinatorial stochastic processes over biologically meaningful discrete random graphs and study their weak limits.  The talk will aim to present the definitions needed to appreciate the five main propositions in this work.  The basic ideas in the proofs will also be presented, if time permits. The mathematical details of the proofs will be presented by the co-author Amandine Veber in the Probability and Analysis Seminar on Thursday September 14th at 1300 hours in 64119, Ångström.

Co-authors: Bhalchandra Thatte (Universidade Federal de Minas Gerais, Belo Horizonte, Brasil) and Amandine Véber (Ecole Polytechnique, Palaiseau, France).

19 September
Speaker: Petter Holme, Suzukakedai Campus, Tokyo Institute of Technology
Title: Spreading processes on temporal networks
Place: Å4003
Time: 12:15 -- 13:00
The contact structure between people affects spreading processes. Network theory has become a powerful tool to understand that relationship. Sometimes, one has information not only about which nodes that interact, but also when the interaction happens. Just like the network topology, structures in time can be very important for the behavior of spreading processes. To leave out the temporal structure could be fatal both for prediction and mechanistic understanding. I will discuss the theory of temporal networks—integrating information about time and network topology. This theory, it turns out, becomes rather different from static network theory (partly because temporal networks are not transitive, in the algebraic sense). I will use disease spreading on temporal networks as my main example, but also discuss the state of the ­field in general, and its future challenges.

3 October
Speaker: Sara Merino Aceituno, Department of Mathematics, Imperial College London
Title: Coupled Self-Organized Hydrodynamics and Stokes models for suspensions of active particles
Place: Å4003
Time: 12:15 -- 13:00
We derive macroscopic dynamics for collective motion in a fluid.The starting point is a coupled Vicsek-Stokes system. The Vicsek modeldescribes self-propelled agents interacting through alignment. It provides aphenomenological description of steric interactions between agents at highdensity. Stokes equations describe a low Reynolds number fluid.

17 October
Speaker: Joakim Stenhammar, Department of Physical Chemistry, Lund University
Title: Correlations and collective behaviours in suspensions of swimming microorganisms
Place: Å4003
Time: 12:15 -- 13:00
Many microorganisms, such as bacteria and algae, have developed swimming strategies to aid them in search for food in their aqueous environments. From a hydrodynamic perspective, these organisms are interesting since swimming on the micron scale is radically different from in the macroscopic world, due to the absence of fluid inertia. Two intriguing hydrodynamic phenomena occurring in suspensions of swimming microorganisms are the transition to so-called ``bacterial turbulence'', whereby bacterial suspensions of sufficiently high density exhibit large-scale flocks and coherent flows, and the enhanced diffusion of passive tracer particles compared to the expected Brownian value. In this talk, I will present results obtained from large-scale numerical simulations of a simple model of hydrodynamically interacting microswimmers that capture both these phenomena accurately enough to draw quantitative conclusions. The results will furthermore be compared to analytical predictions from a stochastic kinetic theory of microswimmer suspensions that goes beyond mean-field, and thus takes into account swimmer-swimmer correlations, showing their importance even far below the transition to bacterial turbulence.

18 October
Speaker: Oleg Iliev,  Fraunhofer Institute for Industrial Mathematics, ITWM
Title: On computer simulation of multiscale processes in porous electrodes of Li-ion batteries
Place: Å4003
Time: 12:15 -- 13:00
Li-ion batteries are widely used in automotive industry, in electronic
devices, etc. In this talk we will discuss challenges related to the
multiscale nature of batteries, mainly the understanding of processes in
the porous electrodes at pore scale and at macroscale. A software tool
for simulation of isothermal and non-isothermal electrochemical
processes in porous electrodes will be presented. The pore scale
simulations are done on 3D CT images of porous electrodes, or on computer
generated 3D microstructures, which have the same characterization as
real porous electrodes. Finite Volume and Finite Element algorithms for
the highly nonlinear problems describing processes at pore level will be
shortly presented. Model order reduction, MOR, empirical interpolation
method, EIM-MOR algorithms for acceleration of the computations will be
discussed, as well as the reduced basis method for studying parameters
dependent problems. Next, homogenization of the equations describing the
electrochemical processes at the pore scale will be presented, and the
results will be compared to the engineering approach based on Newman’s
1D+1D model. Simulations at battery cell level will also be addressed.
Finally, the challenges in modeling and simulation of degradation
processes in the battery will be discussed and our first simulation
results in this area will be presented.

This is joint work with A.Latz (DLR), M.Taralov, V.Taralova, J.Zausch,
S.Zhang from Fraunhofer ITWM, Y.Maday  from LJLL, Paris 6 and Y.Efendiev
from Texas A&M.

31 October
Speaker: Thomas Smed, Company Specialist, Plant Dynamics, Engineering Support Staff, Forsmark Kraftgrupp AB
Title: There is nothing as practical as good theory
Place: Å4003
Time: 12:15 -- 13:00
The seminar will give examples on how linear algebra and system identification is applied to practical problems in Forsmark Nuclear Power Plant.

In order to introduce the subject and explain the context, the seminar will begin with a micro-crash course on control of a BWR.

14 November
Speaker: Attila Szilva, Department of Physics and Astronomy, Uppsala University
Jérôme Michaud, Department of Sociology, Uppsala University
Title: The Voter model with recurrent mobility and Stockholm voting behaviour
Place: Å4003
Time: 12:15 -- 13:00
In order to model the opinion dynamics of voting behaviour in the region of Stockholm, we would like to adapt the Social Influence with Recurrent Mobility (SIRM) variation of the Voter Model to the voting behaviour in the Stockholm region. The initial formulation of the SIRM model has some issues that should be adressed before applying this model to the multiparty situation of Stockholm county. In this talk, we will present the problem and the data we have as well as preliminary results from simulations. For instance, we will introduce a generalized version of the SIRM model that does no suffer from the issues mentioned above.

We will list a few challaneges that could be in the scope of a project of a PhD student. One can study the influence of the commuting network structure on the long time behaviour of the system depending on the parameter of the model. The impact of the network topology on the spatial correlation can be also studied and compared them with real data. As a next step, the underlying social network in the SIRM model could be also rewritten.

28 November
Speaker: Yevgen Ryeznik, Department of Mathematics, Uppsala University
Title: Optimal Designs and Adaptive Randomization Techniques in Clinical Trials
Place: Å4003
Time: 12:15 -- 13:00
Clinical trials are prospective biomedical or behavioral research studies on human subjects that are designed to answer specific questions about biomedical or behavioral interventions. The standard approach of performing such studies is based on the equal allocation of involved subjects between treatment arms. However, it might be that this approach is a nonoptimal choice. In this case, an optimal design and a well-chosen randomization procedure can improve the trial's efficiency. In my talk, I will give a brief overview of the state of the art of the optimal strategies that could be used in real clinical trials. 

5 December
Speaker: Dr. Sturrock, Department of Physiology, Royal College of Surgeons in Ireland
Title: The influence of nuclear compartmentalisation on stochastic dynamics of self-repressing gene expression
Place: Å4003
Time: 12:15 -- 13:00
Short bio:
Dr. Sturrock is a Lecturer in Computational Biology in the Department of Physiology at the Royal College of Surgeons in Ireland. He received his B.Sc. in Applied Mathematics from the University of Dundee in 2009 and then stayed on to complete his PhD under the supervision of Prof. Mark Chaplain in 2013. His thesis research examined spatio-temporal models of gene regulatory networks containing negative feedback loops. Dr. Sturrock then completed a Post Doctoral Research Fellowship at the Mathematical Biosciences Institute within The Ohio State University. Here he worked on projects in macromolecular crowding under the mentorship of Prof. Radek Erban and cell polarization under the mentorship of Dr. Adriana Dawes. Finally, he completed a second postdoc at Imperial College London within a synthetic biology group led by Prof. Mark Isalan where he worked on projects in emergent gene expression and synthetic Turing pattern formation.
Gene expression is an inherently noisy process. This noise is generally thought to be deleterious as precise internal regulation of biochemical reactions is essential for cell growth and survival. Self-repression of gene expression, which is the simplest form of a negative feedback loop, is commonly believed to be employed by cellular systems to decrease the stochastic fluctuations in gene expression. When there is some delay in autoregulation, it is also believed that this system can generate oscillations. In eukaryotic cells, mRNAs that are synthesised in the nucleus must be exported to the cytoplasm to function in protein synthesis, whereas proteins must be transported into the nucleus from the cytoplasm to regulate the expression levels of genes. Nuclear transport thus plays a critical role in eukaryotic gene expression and regulation. Some recent studies have suggested that nuclear retention of mRNAs can control noise in mRNA expression. However, the effect of nuclear transport on protein noise and its interplay with negative feedback regulation is not completely understood. In this talk, I will present four different simple models of gene expression. Through the use of stochastic simulations and applying the linear noise approximation to the corresponding chemical master equations, I will present an investigation of the influence of nuclear import and export on noise in gene expression in a negative autoregulatory feedback loop. I will first present results consistent with the literature, i.e., that negative feedback can effectively buffer the variability in protein levels, and nuclear retention can decrease mRNA noise levels. I will then go on to show that when negative feedback is combined with nuclear retention, an amplification in gene expression noise can be observed and is dependent on nuclear translocation rates. Finally, I will end with a discussion of the effect of nuclear compartmentalisation on the ability of self-repressing genes to exhibit stochastic oscillatory dynamics.

12 December
Speaker: Markos Katsoulakis, University of Massachusetts Amherst, Mathematics & Statistics
Title: Information Inequalities and model-form  Uncertainty Quantification
Place: Å4003
Time: 12:15 -- 13:00
In this talk we discuss information theoretic uncertainty quantification methods for probabilistic models in materials, catalysis and related design problems. We first demonstrate how new, tight and scalable information inequalities can provide computable uncertainty quantification (UQ) indices for observables of interest; these indices account for model-form uncertainty in a neighborhood of a baseline model defined via information divergences. We apply these tools in two  problems arising in catalysis and energy research: a) fast sensitivity screening of complex reaction networks with thousands of parameters and  b) quantifying the impact of multiple sources of uncertainty in mesoscale bayesian reaction networks that include electronic structure data. In the latter context and time permitting,  we will also discuss new UQ methods for probabilistic graphical models.