LIAM Seminar

Can search queries and other novel data streams be utilized for now-casting / forecasting influenza outbreaks? A systematic review of the literature.

Abstract: Influenza is a communicable disorder, which imposes a severe societal and epidemiological burden. However, traditional flu surveillance systems, such as the Centers for Disease Control and Prevention’s (CDC) influenza-like illnesses reports, lag behind real-time by 1-2 weeks. Infodemiology (a port-manteau of “information” and “epidemiology”, coined by the researcher Gunther Eysenbach) and infoveillance (a combination of the words “information” and “surveillance”) represent an emerging approach, enabling the study of the determinants of the digital behavior. Novel data streams (NDSs), including Facebook, Twitter and other social networks posts, as well as Wikipedia access logs, appear to be optimal devices for tracking web activities related to health topics and may complement traditional surveillance systems. Among NDSs, Google Trends (GT) represents a unique open-access tool for monitoring search queries. From 2008 to 2015, Google Flu Trends (GFT), based on GT, has served as a provider of influenza activity estimates for more than 25 countries, until, due to methodological criticisms, GT has decided to stop this project. The present systematic review envisages the feasibility of using NDSs for now-casting/forecasting influenza activity, especially focusing on the recent technical/mathematical advancements in improving the accuracy of search queries-based predictive models of influenza outbreaks.

Dr. Nicola Luigi Bragazzi
Post-Doctoral Fellow
York University

Date: November 19, 2019
Time: 10:30-11:30 AM
Location: Kinsmen 277

Calibration of a dynamic microsimulation of a disease history model for Zika virus (ZIKV) infection

Abstract: While generally causing self-limited and mild illness, Zika virus (ZIKV) infection during pregnancy can cause congenital infection of the fetus leading to features of congenital Zika syndrome (CZS), including microcephaly; and rarely infection of children and adults can cause significant morbidity including congenital Zika syndrome, newborn Zika syndrome, and Guillain-Barré syndrome (GBS). However, our understanding of the population-level impact of the disease on longer-term health outcomes is limited. This presentation describes the development and calibration of a ZIKV disease history model and quantifies the population-level burden of disease of ZIKV for Colombia, using a stochastic microsimulation model.

Method: We developed a ZIKV disease history model using a stochastic, individual-level microsimulation approach. We simulate a dynamic cohort of uninfected individuals over one year using data from the published literature and local census data to parameterize the model and surveillance data from Colombia to calibrate the model. We calibrated five parameters: infection risk, and probabilities of: pregnancy, neurological complications, GBS, and CZS using the Nelder-Mead method and evaluated goodness of fit using the sum of squared differences between simulated proportions and desired proportions.

Result: For the best-fitting parameters the goodness of fit obtained was 6.79×10-10, the incidence of ZIKV infection was 220 per 100,000 people (222 observed), CZS was 4 cases per 1,000 ZIKV infected people (6 observed), neurologic syndrome was 9 cases per 1,000 ZIKV infections (6 observed) and GBS was 9 per 1,000 ZIKV infections (4 expected).

Dr. Raphael Ximenes
Post-Doctoral Fellow
University of Toronto (UofT)

Date: November 12, 2019
Time: 10:30-11:30 AM
Location: Kinsmen 277

Sanofi-York Industrial Research Chair (IRC) Team developments in contact pattern modelling

Abstract: To facilitate the infection process, a contact between a susceptible individual with an infectious agent must take place. In general, a contact refers to either direct physical contact (e.g., person-to-person, droplet spread from cough or sneeze, etc.) or indirect contact (interaction with contaminated food, water, insects, etc.).  Many infectious diseases spread through person-to-person contact and we focus on this transmission route. In this work, we focus on quantifying the contact frequencies between age groups within households in Ontario, Canada. To generate this age-specific contact network, we make use of readily-available demographic information and an established computational approach based on the simulation of a virtual society. In addition, we have developed an alternative method for quantifying contact frequencies between household members using probability theory, specifically using conditional probability functions. We outline these two methods for calculating contact frequencies, as well as the the bigger picture of the application of identifying an age-specific contact network. Joint work with Sanofi-York Industrial Research Chair (IRC) Team.

Zachary McCarthy
PhD Candidate
York University

Date: November 5, 2019
Time: 10:30-11:30 AM
Location: Kinsmen 277

Evolutionary Game Theory and Honey Bees Genetics

Abstract: In this talk, I will present some basic concepts of game theory. Game theory is a process of modelling the strategic interactions among players in a situation containing a set of rules and outcomes. It is a mathematical approach to decision making in conflict situations to understand the optimal solutions to these interactions. It has a wide range of applications, e.g. economics, politics, animal behaviors, business and evolutionary biology etc. My focus would be on one of its form Evolutionary Game Theory (EGT) to evolving populations in the biological context. EGT is connected with the concept of natural selection, which deals with selection and mutation. One of the most important mathematical models of selection comes from evolutionary game theory is replicator dynamics. The replicator dynamics is an ordinary differential equation, which is used to measure the change in the composition of the population over time. In the end, I will discuss the honey bees reproduction system and how can we implement this theory to construct a model on honey bees genetics.

Bushra Majeed
PhD Candidate
York University

Date: October 29, 2019
Time: 10:30-11:30 AM
Location: Kinsmen 277

Decision making of the population under media effects and spread of influenza

Abstract: Media plays a vital role in controlling The decision making of the population in an epidemic.  Influenza, a disease that affects every age group and causes mortality in many cases, is always a concern for public health. The role of mass-media reports on Influenza is studied in Jane. H et al. In this paper, the authors employed a stochastic agent-based model to provide a quantification of mass media reports on the variability in crucial public health measurements. In this research, we are studying the decision making of the population under the effects of media reports and how their decisions contribute to the control of the influenza epidemic. For this purpose, we use a game theory approach to quantify the decision making of the population under the effect of disease and mass media reports. We adapted the model given by Jane. H et al. and modified it by adding sub-compartments to susceptible and vaccinated compartments. We study the movement of the population between these sub-compartments under the media effects in the presence of risk of getting an infection. And how these movements contribute to the incidence rate.

Dr. Safia Athar
Post-Doctoral Fellow
York University

Date: October 16, 2019
Time: 12:30-1:30 PM
Location: Kinsmen 277

Asymptotic Behavior of Abstract Differential Equations with State-dependent Delay

Abstract: In this seminar we are going to talk about some of the goals of my PhD project and present some results on existence, uniqueness of solution and well-posedness for a general class of abstract integro-differential equations with state-dependent delay.

Denis Fernandes
Visiting PhD Student
York University

Date: October 1, 2019
Time: 10:30-11:30 AM
Location: Kinsmen 277

Normalization of periodic delays in delay differential equations arising from population dynamics

Abstract: In this talk, we investigate a scalar delay differential equation with time-varying delay which captures the dynamics of a single species population such as ticks. We consider the possibility of transforming the equation to a delay differential equation with a constant delay. Then we introduce theoretical results on the stability threshold of the scalar linear periodic delay differential equation with numerical examples.

Dr. Kyeongah Nah
Post-Doctoral Fellow
York University

Date: May 14, 2019
Time: 12:00-1:00 PM
Location: Kinsmen 277

A simple review of HSV-2 mathematical models

Abstract: Genital herpes is one of the most prevalent sexually transmitted infections in the world. It is mainly caused by herpes simplex virus type 2 (HSV-2), which is transmissible through skin lesions and mucosa. In this talk, I will give a brief introduction for HSV-2, such as the transmission, symptoms and possible complications, and then present a simple review of the previous studies related to HSV-2.

Qian Li
Visiting PhD Candidate
Xi’an Jiaotong University

Date: Apr 30, 2019
Time: 12:00-1:00 PM
Location: Kinsmen 277

Numerical bifurcation analysis of structured population models

Abstract: Via pseudospectral discretization, a nonlinear delay equation can be approximated with a system of ODEs, and its dynamical and bifurcation properties can be studied with existing software for ODEs. The pseudospectral discretization method can be applied to nonlinear integral and delay-differential equations, with discrete and distributed delays, finite and infinite delays, and some kinds of state-dependent delays.

I will present the method and illustrate its effectiveness with some examples of structured populations from ecology and epidemiology.

Dr. Francesca Scarabel
Post-Doctoral Fellow
York University

Date: Apr 23, 2019
Time: 12:00-1:00 PM
Location: Kinsmen 277

Absenteeism Impact on Local Economy during a Pandemic via Hybrid SIR Dynamics

Abstract: In this paper, we study the cost of absenteeism and presenteeism (going to work while sick) during a pandemic in a local economy with several geographically distinct locations, and with work force populations consisting of individuals who live and work in the same city, and individuals who live and work in different locations (daily commuters).

We run simulations to study the effects of the fear factor and of the severity of disease on the number of missed work days in the region, which we translate into loss of productivity costs. We find that  higher values of the fear parameter lead to high absenteeism and lower infection levels. However, we also show that for severe pandemics (such as the number of secondary infections is higher) there are scenarios where there exists a unique value of the fear parameter which leads to minimum economic costs for the regional economy. This indicates that “staying at home” policies during an epidemic could be implemented for the work force, without reaching a state of emergency level.

Dr. Safia Athar
Post-Doctoral Fellow
York University

Date: Apr 16, 2019
Time: 12:00-1:00 PM
Location: Kinsmen 277

Non-negative matrix factorization and its application on images

Abstract: NMF is a widely used multivariate analysis method on many fields such as face recognition, text clustering, speech processing, blind source separation, image retrieval, and object tracking, and so on. INMF(Incremental NMF) uses incremental learning to lower the computational complexity of NMF when dealing with online processing of large scale data. We proposed OINMF (Incremental NMF with Orthogonal constraints) to improve the efficiency and strengthen the ability of local representation.

Dr. Liuhong Luo
Visiting Professor
Beijing Forestry University

Date: Apr 9, 2019
Time: 12:00-1:00 PM
Location: Kinsmen 277

Public opinion analysis

Abstract: Public opinion analysis is greatly useful for government since there are a lot of hot events and emergency events around us every day. This talk will introduce two application demo systems which are public opinion analysis in China and comparison analysis between global and Chinese public opinion. Then, in this talk I will introduce some epidemic models on public opinion, which are SFI model for one Weibo, SRS/I model for one topic, and some other models for further research.

Dr. Fulian Yin
Visiting Professor
Communication University of China

Date: Mar 26, 2019
Time: 12:00-1:00 PM
Location: Kinsmen 277

Finite-time Control for Time-delay Systems

Abstract: In this talk, we study the finite-time control for time-delay systems. Some linear matrix inequalities (LMIs) based sufficient conditions of finite-time stability (FTS) for neural networks are proposed via designing the controller. And then an example is given to show the effectiveness of our proposed theoretical result. In addition, we develop the Lyapunov-Razumikhin method to FTS and finite-time contractive stability (FTCS) of linear time-varying (LTV) time-delay system. Several sufficient conditions for establishing these FTS properties are obtained. The efficiency of the proposed criteria is illustrated by two numerical examples.

Xueyan Yang
Visiting Student
Shandong Normal University

Date: Mar 19, 2019
Time: 12:00-1:00 PM
Location: Kinsmen 277

IPM Strategies to a Discrete Switching Predator-Prey Model Induced by a Mate-Finding Allee Effect

Abstract: This talk we propose a discrete switching predator-prey model with a mate-finding Allee effect, where also switches are guided by Allee effect.  In this talk, we first study analytically the dynamic behaviors of the two subsystems and the equilibria and their stability of the switched system. Then we provide numerical bifurcation analyses for the switched discrete system. These show that the switched discrete system may have very complex dynamics by 2-parameter bifurcation diagrams which divide the space into regions and study equilibria, and 1-dimensional bifurcation diagrams which reveal that the system has periodic, chaotic solutions, period doubling bifurcations and so on. Furthermore, we try to refer the key parameters and initial densities of both populations associated with pest outbreaks and study their biological implications.

Dr. Wenjie Qin
Visiting Scholar
Three Gorges University

Date: Mar 5, 2019
Time: 12:00-1:00 PM
Location: Kinsmen 277

Approximate Controllability of Second-order Semilinear Evolution Systems with Finite Delay

Abstract: In this work, we study the approximate controllability for a class of semilinear second order control systems with finite delay. Sufficient conditions for approximate controllability are established by constructing fundamental solutions and using the resolvent condition and techniques on cosine family of linear operators. To illustrate the applications of the obtained results, an example is provided in the end.

Xiaofeng Su
Visiting PhD Candidate
East China Normal University

Date: Feb 26, 2019
Time: 12:00-1:00 PM
Location: Kinsmen 277

Convolutional Autoencoders for Light Field Compression

Abstract: With computing power, storage and bandwidth all constantly growing, file compression is and will continue to be an inevitable constraint on new technology. 3D video and immersive content are on the horizon, but the size of the data continues to be a large problem in application.

We aim to give a overview of one such media format known as light field images. We will also discuss possible solutions for the encoding and decoding of such data in an attempt to work towards a full pipeline of an immersive file transfer protocol. In particular we will highlight the idea of deep neural networks, and use the autoencoder network structure in combination with convolutional filters to achieve a data compression in terms of total data points. Lastly, we will touch on some of the details of the actual training of the network itself.

Zarko Valtchev
PhD Candidate
York University

Date: Feb 5, 2019
Time: 11:45-12:45 PM
Location: Kinsmen 277

Liam in 2019- where we are heading for?

Dr. Jianhong Wu
University Distinguished Research Professor
Canada Research Chair in Industrial and Applied Mathematics
Laboratory for Industrial and Applied Mathematics (LIAM)
Department of Mathematics and Statistics
York University, Toronto, Canada

Date: Jan 22, 2019
Time: 12:00-1:00 PM
Location: Kinsmen 277

Invariant bundles of positive linear cocycles

Abstract:  It is known that solutions of an autonomous differential equation form a one-parameter group that defines a phase semiflow. Analogous to autonomous cases, non-autonomous equations give rise to a skew product semiflow. In fact, linear cocycles are the discrete version of linear skew product semiflows.

For an autonomous ordinary differential equation, eigenvalues and eigenvectors of the generator of its corresponding linear equation are a main tool to describe its local behavior around an equilibrium. In a sense, invariant bundles can be consider as a much more general version of eigenvectors, which describe the local behavior along a non-trivial solution of autonomous (or non-autonomous) evolution equations.

In this talk, I will introduce the method to obtain the existence of compact positive linear cocycles and problems still haven’t been solved in the non-compact case.

Dr. Lirui Feng
Post-doctoral fellow
York University

Date: Jan 15, 2019
Time: 12:00-1:00 PM
Location: Kinsmen 277

Periodic Solutions of Abstract Semilinear Equations with Applications to Biological Models

Abstract: We study the existence of periodic solutions to the abstract semilinear evolution equationdu/dt=A(t)u(t)+F(t,u(t)), t \geq 0in a Banach space X, where A(t) is a T-periodic linear operator on X (notnecessarily densely defined)  satisfying the hyperbolic conditions, and Fis continuous and T-periodic in t. The idea is to combine Poincare maptechnique with fixed point theorems to derive various conditions on theoperator A(t) and the map F(t, u) to ensure that the abstract evolutionequation has periodic solutions.

Three cases are considered: (i) If A(t)=Ais time-independent and is a Hille-Yoshida operator, conditions on F are given to guarantee the existence of mild periodic solutions; (ii) If A(t) is time-dependent and satisfies the hyperbolic condition, sufficient conditions on A(t) and F are presented to ensure the existence of mild periodic solutions; (iii) If A(t)=A is time-independent, is a Hille-Yoshida operator and generates a compact semigroup, the existence of mild periodic solutions is also discussed.

As applications, the main results are applied to establish the existence of periodic solutions in a delayed periodic red-blood cell model; age-structured models with periodic harvesting, diffusive logistic equations with periodic coefficients, and periodic diffusive Nicholson’ blowflies equation with delay.

Dr. Qiuyi Su
Post-doctoral fellow
York University

Date: Dec 4, 2018
Time: 10:30-11:30 AM
Location: Kinsmen 277

Impact of influenza vaccine modified effects on the outcomes of immunization in a population

Abstract: Flu vaccines are designed to provide protection against flu infection and its complications.
Evidence from experimental and observational studies suggest that flu vaccines may also influence the intensity and the duration of infectiousness of vaccinees. The above-mentioned vaccine modified effects (direct effects) are the characteristic of the vaccine product or the receiver which are independent of the population and the vaccination program. In comparison, the vaccine effects estimated from the observational studies are the combined outcomes of the direct effect and the indirect effect of an immunization program.

Dr. Kyeongah Nah
Post-doctoral fellow
York University

Date: Nov 27, 2018
Time: 10:30-11:30 AM
Location: Kinsmen 277

Linking influenza infection to risk of severe cardiovascular events: a mechanistic approach.

Abstract: Influenza poses major threats to public health worldwide. This seminar talk will outline the basic biological pathways linking in influenza infection to cardiovascular events such as myocardial infarction. The primary pathways related to cardiovascular events include the immune response, inflammatory response, and blood coagulation cascade. Recently, mathematical models have been constructed in an effort to describe these biological system dynamics. A brief review of existing mathematical modelling, experimental works, as well as data limitations for informing these models will be presented. Finally, the York-Sanofi in-host modelling team’s progress in this area as well as our future directions will be outlined.

Zachary McCarthy
PhD Candidate
York University

Date: Nov 20, 2018
Time: 10:30 AM-11:30 AM
Location: Kinsmen 277

Recent progress on the propagation phenomena in a shifting habitat

Abstract: In this talk, I will first introduce the biological motivation and background on this topic.  Then I will report the main results in the paper, written by H. Berestycki and J. Fang, “Forced waves of the Fisher–KPP equation in a shifting environment”.  At last, I will briefly discuss about the mathematical definition of rotating waves for scalar reaction-advection-diffusion equations on a circle, the possible connection with forced waves in a shifting habitat and the preliminary results in my mind.

Dr. Xiao Yu
Post-doctoral fellow
York University

Date: Oct 23, 2018
Time: 10:30 AM-11:30 AM
Location: Kinsmen 277

Modelling and Analyzing Virus Mutation Dynamics of Chikungunya Outbreaks

Abstract: Chikungunya fever, caused by chikungunya virus (CHIKV) and transmitted to humans by infected Aedes mosquitoes, posed a global threat during the outbreaks in several countries of the Americas in 2015. Recent evidence from La Reunion, Italy and China indicates the existence of a new variant of CHIKV with a shorter extrinsic incubation period in contaminated mosquitoes, but the role of this new variant on the spread of chikungunya fever locally and globally is unclear.
Here, we develop a compartmental model to address the virus mutation dynamics, by assuming a linear virus mutation rate among contaminated mosquitoes. Preliminary numerical simulations of the model show that a substantial virus mutation rate, combined with high virus transmission probabilities from mosquito to human, could result in sustainable chikungunya fever outbreaks.
We apply Markov chain Monte Carlo sampling method to test our model to the 2007 chikungunya fever outbreak data in Italy where the mutant strain was discovered. We conclude that the basic reproduction number might be underestimated without considering the mutation dynamics, and our estimation shows that the basic reproduction number of the 2007 Italy outbreak was R0 = 2:035 [95%CI : 1:9424 – 2:1366]. Sensitivity analysis shows that the transmission rate of the mutant strain from mosquitoes to human is more influential on R0 than the shortened extrinsic incubation period. We conclude that the virus mutation dynamics could play an important role in the transmission of CHIKV, and there is a crucial need to better understand the mutation mechanism.

Dr. Xiaomei Feng
Lecturer of Applied Mathematics
Yuncheng University, China

Date: Oct 9, 2018
Time: 10:30 AM-11:30 AM
Location: Kinsmen 277

An Application of the Projective Clustering Algorithm PART to Sports Analytics

Abstract: We aim to show how a neural network based machine learning projective clustering algorithm, Projective Adaptive Resonance Theory (PART), can be effectively used to provide data-informed sports decisions. In this case, we are trying to overtake the sports director role of AS Roma Monchi by studying and merging together two databases: AS Roma players and players linked with possible transfer moves to AS Roma in Summer 2018. Each player is viewed as a vector and its dimensions are the 47 abilities of the players according to a soccer database. This is a high dimensional data as players should be grouped only in terms of their performance with respect to a small subset of attributes. Projective clustering analyses provide a purely data-driven analysis to identify critical attributes and attribute characteristics for a group of players to form a natural cluster (in lower dimensional data space) in an unsupervised way. By merging the two databases, our unsupervised clustering analysis provides evidence-based recommendations about the Club team formation, and in particular, the decision to buy and sell players within the same clusters, under different scenarios including financial constraints.

Marco Tosato
PhD Candidate
York University

Date: Sep 18, 2018
Time: 10:30 AM-11:30 AM
Location: Kinsmen 277

Modelling of meningococcal disease dynamics during mass gathering in the Hajj

Abstract: The spread of infectious diseases in mass gathering (MG) setting is considered a major public health threat requiring a unique preparedness and implementation of effective preventive measures. The convergence of massive number of people in a temporally and spatially narrow dimension is understandably associated with high likelihood of infectious disease spread, and thus the understanding of associated risk of disease transmission and outbreak is crucial for public health preparedness. We propose a model framework for mass gathering and present an application to meningococcal disease during the Hajj by studying an SVCIR deterministic metapopulation model with residency in order to understand the disease dynamics due to force of transmission and assess the outbreak risk.

Kazi Rahman
Postdoctoral fellow
York University

Date: Aug 14, 2018
Time: 10:30 AM-11:30 AM
Location: Kinsmen 277

Disease Dynamics and Vaccine implementation for Meningococcal Disease during Hajj Pilgrimage

Abstract: Millions of Muslims migrate to Mecca in Saudi Arabia every year for a religious pilgrimage called ‘Haji’. This mass gathering is one of largest in the world and draws people from various countries. The high density of people in this region creates an environment which facilitates health problems, including the transmission of pathogens, many of which cause frequent epidemics in this region and has shown to spread abroad when travellers return to their resident countries. A pathogen that has caused frequent epidemics and serious health implications for the Haji pilgrims is the bacterium Neisseria meningitidis. It causes meningococcal disease (MCD), which is leading cause of bacterial meningitis. The transmission mode and the pathogenesis of the disease were examined in this presentation. This presentation also discusses the environmental factors and the bacterium’sgenetic factors which contribute to the transmission and disease dynamics of MCD. This presentation also explains the vaccine implementation in place for MCD during annual Hajj.

Fabian Wong
Visiting student
University of Western Ontario

Date: June 12, 2018
Time: 11:30 AM-12:30 PM
Location: Kinsmen 277

Study of a mathematical model of HIV spread including PrEP

Abstract: Pre-exposure prophylaxis (PrEP) consists in the use of an antiretroviral medication to prevent the acquisition of HIV infection by uninfected individuals and has recently demonstrated to be highly efficacious for HIV prevention. However, taking this treatment is often accompanied by a discontinuation of the use of condoms which causes an increase in the spread of other STIs. I study a model proposed by C. Silva and F. Torres in (Modeling and Optimal Control of HIV/AIDS Prevention ThroughPrEP, 2017) of HIV spread including PrEP treatment. Existence, uniqueness and global stability of equilibria are proved. This model is applied to a Cape Verde case study to show the potential impact of this treatment only on the HIV spread. Finally, I propose ways to enrich the model in order to use it for a new problem, the impact of PrEP on the spread of other STIs.

Thomas Martin
University Claude Bernard Lyon1

Date: June 12, 2018
Time: 10:30-11:30 AM
Location: Kinsmen 277

Data Analysis and Decision Support of Radio and Television

Abstract: In traditional industry chain of radio and television, audience could only passively accept TV information resources and their demand cannot be transmitted to any part of the industrial chain. We analyse the user behavior, design some decision support models like personality recommendation, advertising, program evaluation, and establish visualization systems. After data analysis and decision support, it is possible to get an innovative industry chain and the user experience can be greatly improved.

Fulian Yin

York University

Date: May 15, 2018
Time: 10:30-11:30 AM
Location: Kinsmen 277

Mechanistic movement models to understand epidemic spread

Abstract: An overlooked aspect of disease ecology is considering how and why animals come into contact with
one and other resulting in disease transmission. Mathematical models of disease spread frequently
assume mass-action transmission, justified by stating that susceptible and infectious hosts mix
readily, and foregoing any detailed description of host movement. Numerous recent studies have
recorded, analysed and modelled animal movement. These movement models describe how animals
move with respect to resources, conspecif ics and previous movement directions and have been used
to understand the conditions for the occurrence and the spread of infectious diseases when hosts
perform a type of movement. Here, we summarize the effect of the different types of movement on
the threshold conditions for disease spread. We identify gaps in the literature and suggest several
promising directions for future research. The mechanistic inclusion of movement in epidemic models
may be beneficial for the following two reasons. Firstly, the estimation of the transmission coefficient
in an epidemic model is possible because animal movement data can be used to estimate the rate of
contacts between conspecifics. Secondly, unsuccessful transmission events, where a susceptible host
contacts an infectious host but does not become infected can be quantified. Following an outbreak,
this enables disease ecologists to identify ‘near misses’ and to explore possible alternative epidemic
outcomes given shifts in ecological or immunological parameters.

Abdou Moutalab Fofana
Memorial University of Newfoundland

Date: April 27, 2018
Time: 10:30-11:30 AM
Location: Kinsmen 277

Bayesian Inference for the pricing of Quanto options

Abstract: In our study, the pricing of Quanto options is studied. We show that how one can make inference and compute option prices using Bayesian method, where the underlying asset volatility and the exchange rate volatility and the correlation between both are estimated based on conditional posterior densities. The numerical posterior simulation, performed by implementing the MCMC algorithm, demonstrates that the Bayesian method developed in the paper has its advantages in comparison with the maximum likelihood estimation, especially for in-the-money options.

Lisha Lin
Visiting PhD candidate
York University

Date: April 24, 2018
Time: 10:30-11:30 AM
Location: Kinsmen 277

Piecewise models of immune-virus system

Abstract:To overcome the problems such as adherence difficulties and evolution of drug resistance generated from the continuous treatment of HIV-infected patients, structured treatment interruptions (STIs), as an alternative strategy, have been suggested to be a good choice for some chronically infected individuals who may need to take drugs throughout their lives. The purpose of the study is to propose mathematical models describing the STIs for treating HIV patients, and examine the efficacy of this treatment for controlling the plasma HIV-1 RNA below a certain level and maintaining the activity of the immune system of the patients.

Dr. Biao Tang
Post-doctoral fellow
York University

Date: April 17, 2018
Time: 10:30-11:30 AM
Location: Kinsmen 277

Tick-borne Encephalitis model with co-feeding transmission

Abstract:   We are modeling a tick-borne encephalitis dynamics model including co-feeding transmission. We stratify the host population in terms of attached nymph number and consider grooming behavior, an adaptive response to reduce tick densities. We investigate the effects of grooming behavior, contact rate, and co-feeding transmission on tick density distribution.

Dr. Xue Zhang 
Post-doctoral fellow
York University

Date: March 27, 2018
Time: 10:30-11:30 AM
Location: Kinsmen 277

Exponential separation for non-compact operators

Abstract:  Roughly speaking, exponential separation (abbr. ES) describes the growth rate of linearization of nonlinear dynamical systems. It is understood that ES would be an important tool to analyze local behaviors of systems near by their invariant sets. Moreover, It also has closed connections to many other topics in different disciplines of dynamical systems, such as Dominated splitting in differential systems, Exponential dichotomy in differential equations, Multiple ergodic theorems in ergodic theory, and so on.  Before utilizing it, We will naturally ask if the ES property can be established. Motivated by this question, there are several famous theorems to answer partly. For finite dimensional systems, it is the Perron-Frobenius theorem. For infinite compact dimensional systems, it’s the Krein-Rutman theorem. But in the case of non-compact systems, Can we establish ES property?  I will talk about parts of our works.

Dr. Lirui Feng 
Post-doctoral fellow
York University

Date: March 6, 2018
Time: 10:30-11:30 AM
Location: Kinsmen 277

Risk of Tick-borne Encephalitis transmission in Hungary

Abstract: Tick-borne encephalitis (TBE) is a central nervous system infection which is endemic in many European countries including Hungary. In this study, we estimate the ecological/epidemiological parameters for TBE transmission using climate data and TBE incidence data. With the resulted TBE transmission model, we assess the risk of TBE transmission in Hungary.

Dr. Kyeongah Nah
Post-doctoral fellow
York University

Date: Feb 20, 2018
Time: 10:30-11:30 AM
Location: Kinsmen 277

Inferring epidemiological dynamics of infectious diseases using Tajima’s D statistic on nucleotide sequences of pathogens

Abstract: The estimation of the basic reproduction number is essential to understand epidemic dynamics, and time series data of infected individuals are usually used for the estimation. However, such data are not always available. Methods to estimate the basic reproduction number using genealogy constructed from nucleotide sequences of pathogens have been proposed so far. Here, we propose a new method to estimate epidemiological parameters of outbreaks using the time series change of Tajima’s D statistic on the nucleotide sequences of pathogens. To relate the time evolution of Tajima’s D to the number of infected individuals, we constructed a parsimonious mathematical model describing both the transmission process of pathogens among hosts and the evolutionary process of the pathogens. As a case study we applied this method to the field data of nucleotide sequences of pandemic influenza A (H1N1) 2009 viruses collected in Argentina. The Tajima’s D-based method estimated basic reproduction number to be 1.55 with 95% highest posterior density (HPD) between 1.31 and 2.05, and the date of epidemic peak to be 10th July with 95% HPD between 22nd June and 9th August. The estimated basic reproduction number was consistent with estimation by birth–death skyline plot and estimation using the time series of the number of infected individuals. These results suggested that Tajima’s D statistic on nucleotide sequences of pathogens could be useful to estimate epidemiological parameters of outbreaks.

Dr. Kiyeon Kim
Visitor Scholar
York University

Date: Feb 6, 2018
Time: 10:30-11:30 AM
Location: Kinsmen 286

An agent-based modelling (ABM) tool for forced migration scenario simulation

Abstract: We present an agent based modelling tool which could be used to address complex reasons for movements of people, particularly in context of mixed movements (e.g. economic factors, conflict, human rights violations/persecution/torture, environmental change, human trafficking etc.) in order to be better prepared for assisting and protecting displaced populations. This research is a part of the main project “Building Bridges across Social and Computational Sciences: Using Big Data to Inform Humanitarian Policy and Interventions”. If interested, more details can be found at http://fmbd.info.yorku.ca/

Dr. Kazi Rahman
Post-doctoral fellow
York University

Date: Jan 30, 2018
Time: 10:30-11:30 AM
Location: Kinsmen 286

Dynamical behaviors of antimicrobial continuation and de-escalation models

Abstract: In our previous modelling work, “Benefits and unintended consequences of antimicrobial de-escalation: implications for stewardship programs”, we observed that de-escalation can be beneficial in terms of reducing strain transmissions under certain parameter settings. However, due to the complexity of the model, we were not able to mathematically characterize the impacts of these parameters on the model dynamics. In this talk, I will show some recent mathematical analysis of two simplified models of antimicrobial de-escalation and continuation, so as to better explain and further understand our prior results.

Dr. Xi Huo
Assistant Professor
University of Miami

Date: Jan 9, 2018
Time: 10:30-11:30 AM
Location: Kinsmen 286

Self-Excited Periodic and Quasi-Periodic Vibrations for Higher
Dimensional Damped Wave Equations.

Abstract: Using techniques from local bifurcation theory, we prove the
existence of various types of temporally periodic and quasi-periodic
waves for damped wave and beam equations, in higher dimensions. The
emphasis is on understanding the role of external bifurcation
parameters and symmetry, in generating the periodic/quasi-periodic
motion. Most of the work presented is joint with Brian
Pigott.

Dr. Nemanja Kosovalic
Assistant Professor
University of South Alabama

Date: Dec 15, 2017
Time: 11:30 AM-12:30 PM
Location: Kinsmen 286

Estimating reproduction numbers using nucleotide sequence data by the “Bayesian evolutionary analysis by sampling trees”(BEAST)

Abstract: Bayesian phylogenetic methods are commonly used for rapidly mutated viruses, which can affect reconstructed tree structure, to infer epidemiological processes from genetic data. Here I will introduce basic of the Bayesian theorem with BEAST briefly and Birth-death(BD) model which make the BEAST available to estimate reproduction number. After the introduction, I will show examples of its application to real sequence data, pandemic(H1N1) 2009 in Argentina and Hepatitis C virus(HCV) in Egypt.

Dr. Kiyeon Kim
Visitor Scholar
York University

Date: Nov 28, 2017
Time: 10:00-11:00 AM
Location: Kinsmen 286

Study on data mining method of gene transcriptome in tuberculosis

Abstract: This study will utilize the characteristics of transcriptome data related with TB, through significant analysis, association rule analysis, multi-level genetic model and dynamic time warping model to discover more key host gene associated with tuberculosis and their categories; to research the spatial correlation of all kinds of key genes and their roles in the formation of tuberculosis; to analysis dynamic changes of gene transcription on different time scales and the transcription rules in various conditions; to predict the potential expressions of pathogenic genes.

Dr. Xu Zhang
Post-doctoral Fellow
York University

Date: Nov 21, 2017
Time: 10:00-11:00 AM
Location: Kinsmen 286

Agent-Based model development for with-in host dynamics of L. monocytogenes

Abstract: The case fatality and illness rates associated with L. monocytogenes remain unchanged over the decades despite the significant efforts and control protocol obtained by private and public sectors.In order to demonstrate the human gastro-intestinal pathway of L. monocytogenes, we develop an agent based model. I will demonstrate the impact of food intake pattern, stomach pH variation and storage condition on the survival of L. monocytogenes in the stomach. The model will also illustrate the role of immune potential to prevent intestinal infections.

Dr. Ashrafur Rahman
Post-doctoral Fellow
York University

Date: Nov 14, 2017
Time: 10:00-11:00 AM
Location: Kaneff Tower

Sleep Duration and Chronic Condition Among Canadian Adults: Do Mental Illness Play a Mediating Role?

Abstract: Chronic condition has been major contributors to reduced quality of life, loss of productivity, and increased hospitalization and health care costs as well as premature death in Canada.For better chronic condition prevention, this large-scale study was designed to explore the potential association of sleep duration with chronic condition and mediation by mental illness. We obtained data from the 2011-2012 Canadian Community Health Survey. A total of 40,614 participants aged 18 years or older from four provinces (Nova Scotia, Quebec, Manitoba, and Alberta) that participated in the sleep survey module were selected for the study. Logistic regressions were performed to assess the mediation of mental illness on the association between sleep duration and chronic condition. The age- and sex- standardized prevalence of any chronic condition in four provinces of Canada was 54.5%. Compared to those sleep 7 to < 9 h, participants in both short (< 5 h, and 5 to < 7 h) and long (9 to < 11 h, and ≥ 11 h) sleep duration reported a higher prevalence of any chronic condition. After adjusting for all potential confounders, the “U-shaped” association between sleep duration and any chronic condition persisted. Following the criteria of examining mediating effects, mental illness was found significantly mediated the relationships between sleep duration and any chronic condition(all Sobel P< 0.001). However, the mediated effect size of mental illness was obviously higher in long sleep duration (32.8% and 33.4%) than short sleep duration (14.0% and 9.5%). Sleep duration had U-shaped relationships with the presence of chronic condition. Mental illness play a mediating role on the relationships between sleep duration and chronic condition, especially in long sleep duration.

Dr. Haijiang Dai
Post-doctoral Fellow
York University

Date: Nov 7, 2017
Time: 10:00-11:00 AM
Location: Kaneff Tower

Understanding the dynamics of West Nile Virus in Emilia-Romagna, Italy

Abstract:  West Nile Virus (WNV) has been identified for the first time in Italy in 1998, and more continuously since 2008 with a total of 173 neurological human cases between 2008 and 2015. Still the circulation of the virus appears to have been episodic with most cases concentrated in a few years and a few hotspots shifting in different years.  The region Emilia Romagna, which is one of the most affected areas, has set up since 2009 a systematic program of mosquito and corvids (known to be among the most competent bird species for WNV) trapping and testing. Data collected through this program have been analysed through a mathematical model in order to understand the main drivers of the observed dynamics. The analysis has mainly been based on an SIR (for competent birds)-SI (mosquitoes) model, with an environmentally driven population model, validated on independent data, for mosquito dynamics, and a simple population model for bird dynamics, in which the free parameters  were the mosquito biting rate  and the host-vector ratio. Our results showed that simplest models with constant mosquito feeding behaviours are incompatible with the observed seasonal patterns of infected mosquitoes and birds. On the other hand, including a seasonal shift in mosquito feeding behaviour makes model outputs much more consistent with observed data. Our findings can be of particular interest for public health policy makers, as they provide important insights on WNV dynamics in order to improve surveillance, and risk assessment of WNV in the area.

Marco Tosato
PhD Candidate
York University

Date: Oct 24, 2017
Time: 10:00-11:00 AM
Location: Kaneff Tower

pH dependent C. jejuni thermal inactivation models and application to poultry scalding

Abstract: Campylobacter jejuni related outbreaks and prevalence on retail poultry products pose threats to public health and cause financial burden worldwide. To resolve these problems, it is imperative to take a closer look at poultry processing practices and standards. Using available data (D-values) on the thermal inactivation of C. jejuni we develop a comprehensive inactivation model, taking into account the variation of strain-specific heat resistance, experimental method, and suspension pH. Utilizing our C. jejuni thermal inactivation model, we study the poultry scalding process. We present a mechanistic model of bacteria transfer and inactivation during a typical immersion scald
in a high-speed industrial plant. Integration of our 
C. jejuni inactivation model into the scalding model culminates in validation against industrial processing data. In particular, we successfully predict bacteria concentrations in the scald water and link key factors such as scald water pH and temperature to cross-contamination and overall microbiological quality of carcasses. Furthermore, we demonstrate the applicability of our inactivation model for scalding operations at seven Canadian poultry plants. In addition to providing recommendations for best-practice and a review of scalding research, our work is intended to act as a modular foundation for further research in the interest of public health and financial well-being.

Zack McCarthy
PhD Candidate
York University

Date: Oct 24, 2017
Time: 10:00-11:00 AM
Location: CB 126

Lone star tick and its dynamics

Jemisa Sadiku , Zilong Song,
PhD Candidate
York University

Date: Oct 10, 2017
Time: 10:00-11:00 AM
Location: CB 126

Effect of Host Resistance on Tick Population Dynamics  

Abstract: The purpose of this model is to demonstrate the tick population dynamics when we consider bitten host with resistance. The reason why we consider such observation it’s because there is evidence of tick’s population size to be reduced if they are biting hosts with resistance.The reaction of sensitized host when a tick starts feeding is very different from that of susceptible ‘naive’ host. For instance, the feeding site of sensitized host is characterized by erythema and also by intra-epidermal vesicles packed with basophils which will result in immediate death of the tick. On the other hand susceptible hosts have very little and slow reaction to the tick feeding site allowing the tick to fully finish the feeding process and hence to achieve the disease transmission to the host.

In addition, we will observe how the values of parameters are correlated with tick’s population dynamics and also observe if these biological factor are supported by the mathematical model.

Jemisa Sadiku (Joint work with Mahnaz Alavinejad),
PhD Candidate
York University

Date: Oct 3, 2017
Time: 10:00-11:00 AM
Location: CB 126

Bayesian Inference of Multiple Gaussian Graphical Models

Abstract: I will present a Bayesian approach to inference of multiple networks that has been proposed by Stingo et al. (JASA, 2015). The Bayesian approach readily allows the incorporation of crucial features into a model, including sharing of graph structure across related sample groups and providing a means for obtaining a measure of relative network similarity across groups. The approach also provides the the ability to include prior knowledge of edge-specific interactions and to encourage the degree of similarity to an established network. I will give a brief background on Bayesian inference and graphical modeling of network data and then describe the model specification, including the choice of prior distributions in the model by Stingo et al. I will present the numerical methods used for posterior inference and model selection and conclude with simulation results and an application of the methods to modeling protein networks.

Chris Prashad
PhD Candidate
York University

Date: June 19, 2017
Time: 3:00 PM – 4:00 PM
Location: CB 126


A Model of Chikungunya Transmission with Virus Mutation

Xiaomei Feng
Lecturer of Applied Mathematics
Yuncheng University, China

Date: April 17, 2017
Time: 12:30 PM – 1:30 PM
Location: CB 126

Population Dynamics and Genetic Studies of Sex-Determining Alleles in Honey Bees

Abstract:Crossing of different races of honeybees has become a common practice in these days. But many breeders fear that no pure races will be available in future to conserve the local honeybee ecotypes. Therefore population genetic studies are necessary to establish the expected viability of brood in such populations. Allele frequencies are responsible for viability and is used to characterize the genetic diversity in population. In this talk I will present some work of Jerzy Woyke, ”Population Genetic Studies on Sex Alleles in Honey Bee Using the example of the Kangroo Island Bee Sanctuary”, Journal of Apicultural Research 15(3/4), 1976, in which he formulated and analyzed the mathematical models and analytical expression for calculating the frequencies of sex alleles in subsequent generations and two sexes of honeybees. Also, I will discuss some mathematical theories of the population dynamics of the sex determining alleles in honeybees, developed by ShozoYokoyama and Mastoshi Nei, ”Population Dynamics of Sex-Determining Alleles in Honey Bees and Self-Incompatibility Alleles in Plants”, Genetics 91(3), 1978, where they proved that in an infinitely large populations with n number of alleles, the equilibrium frequency of sex alleles is 1/n and the asymptotic rate of approach to this equilibrium is 2/3n per generation.

Bushra Majeed
M.A Student
York University

Date: April 3, 2017
Time: 12:30 PM – 1:30 PM
Location: CB 126

Network of neurons with delayed feedback and dynamic memory enhancement.

Abstract:Delayed negative feedback, coupled with absolute refractory period, can generated a large amount of stable periodic orbits for associate memory storage and retrieval.

Dr. Jianhong Wu
University Distinguished Research Professor
Canada Research Chair in Industrial and Applied Mathematics
Laboratory for Industrial and Applied Mathematics (LIAM)
Department of Mathematics and Statistics
York University, Toronto, Canada

Date: March 20, 2017
Time: 12:30 PM – 1:30 PM
Location: CB 126

Complementary Sex Determination Substantially Increases Extinction Proneness of Haplodiploid Populations

Abstract: Haplodiploid insects such as ants, bees, and wasps are very important components of terrestrial ecosystems, and their conservation is essential for economic as well as ecological reasons. The conservation genetics of haplodiploids has received very little attention. Haplodiploidy is the property of hymenoptera due to complementary sex determination. In this unique system females develop from the fertilized egg while haploid males are produced from unfertilized egg, however diploid male are also produced but homozygous at the sex-determining locus. In this talk, I will present the work of Amro Zayed and Laurence Packer in which they discussed that the complementary sex determination mechanism in hymenoptera through homozygosity, leading to the production of more sterile and inviable diploid males, due to which haplodiploids are substantially more, rather than less, prone to extinction. Some results on the base of stochastic modelling will also be provided to see the effect of diploid male production (DMP) on the extinction dynamics of haplodiploid populations.

Bushra Majeed
M.A Student
York University

Date: January 30, 2017
Time: 12:30 PM – 1:30 PM
Location: CB 126

Modelling the Evolution of Influenza towards to its Prediction

Abstract: The transmission dynamics of infectious disease is non-linear, since the expected number of transmission events is proportional to the number of both susceptibles and invectives. Mathematical modelling has been playing a key role in understanding and predicting the transmission dynamics. An important disease which is difficult to understand and predict is Influenza. The difficulty in predicting influenza dynamics arises from i). the host of influenza is not only human, the most strains causing pandemic come from the non-human population, but the field data of non-human cases is not enough for predicting the invasion dynamics of influenza from non-human hosts; ii) Even in human population influenza virus evolves quickly and the efficacy of vaccine wanes quickly, so modelling the evolution of influenza in human population is challenging; iii) The transmission probability of influenza is changing
over time, it correlates with the absolute humidity and so the model coefficients in the nonlinear term is a function of time. In this talk, I will introduce some progress how mathematical modelling contributes to the understanding the transmission dynamics of influenza, especially for ii) and iii).

Dr. Ryosuke Omori
Assistant Professor
Research Centre for Zoonosis Control, Hokkaido University, Japan

Date: January 09, 2017
Time: 2:30 PM – 5:30 PM
Location: North Ross 201

Descriptive versus mechanistic dose-response modeling of L. monocytogenes infection in human population

Abstract: Dose-response relationship of L. monocytogenes infection is fundamental in evalution of risk analysis. Descriptive models (exponential, log-logistic and betarpoisson) describing the dose-response relationships have been widely used in L. monocytogenes outbreaks. These models, unfortunately, lack the insights of host-pathogen interaction that drive the response outcomes. Recently, we have developed a mechanistic model to account for the host-pathogen interaction in mouth to gut pathway that provides a mechanistic basis of dose-response relationship for L. monocytogenes infection. Our current study looks into the differences and similarities of two modeling approaches. In particular, we identify the key parameters and their relative ranges that differentiate the models in L. monocytogenes outbreak to human population.

Dr. Ashrafur Rahman
Post-doctoral fellow

Date: December 15, 2016
Time: 10:30 am to 11:30 am
Location: CB 126

The Lyapunov’s functions of some epidemic models of control with vaccine: Case of Tuberculosis and Polio.

Abstract: For the TB, the spread dynamics of the tuberculosis through mathematical models that incorporates both latent and clinical stages has been proposed by HONGBIN GUO and MICHAEL Y. LI.
In this work, we modified the precedent model and we add a strategy of control with vaccine. Then we calculated the basic reproduction number (R0). If R0 ≤ 1, the TB always dies out, otherwise the tuberculosis becomes endemic. The global stability of endemic equilibrium is established through direct Lyapunov and Lasalle Methods. We confirmed our analytics results through numerical simulations.
Talking about Polio, because of the lack of treatment of that disease, the only mean of prevention is immunization through live oral polio vaccine (OPV) or/and inactivated polio vaccine ( IPV).
Poliomyelitis is a very contagious viral infection caused by poliovirus. Children are principally targeted.
In this paper, we assessed the impact of vaccination in the control of spread of poliomyelitis. We used the deterministic SVEIR model of infectious disease transmission (Susceptible-Vaccinated-Latent-Infectious-Removed), where vaccinated individuals are also susceptible even in a lesser degree.
Using Lyapunov-Lasalle methods, we proved the global asymptotic stability of the unique endemic equilibrium whenever Rvac > 1. Some numerical simulations based on poliomyelitis data from Cameroon, conducted us to approve analytic results and demonstrated the importance of vaccinate coverage when it comes to controlling the spread of that disease.
In terms of perspectives, a study related to Polio, of multi group cases with migration between compartments of the same epidemic status is to be issue.

Dr Léontine NKAMBA
Senior Lecturer
University of Yaounde
Higher Teacher Training College
Department of Mathematics

Date: December 8, 2016
Time: 10:30 am to 11:30 am
Location: CB 126

Does two lags really make a trouble for stability analysis?

Seminar Notes

Dr. Jianhong Wu
University Distinguished Research Professor
Canada Research Chair in Industrial and Applied Mathematics
Laboratory for Industrial and Applied Mathematics (LIAM)
Department of Mathematics and Statistics
York University, Toronto, Canada

Date: December 1, 2016
Time: 10:30 am to 11:30 am
Location: CB 126

Investigating the optimal dengue vaccination to mitigate Zika cases

Biao Tang
Visiting PhD Candidate
Xi’an Jiaotong University, China,
York Univerisy and Fields Institution, Toronto, Canada

Abstract: This is a follow-up study of our previous model on implication of dengue vaccination for Zika outbreak, in which we found that due to the antibody-dependent enhancement, vaccinated individuals with antibody against dengue could have higher chance to be contaminated by Zikavirus, thus dengue vaccination could induce more Zika cases. In this study, we perform numerical experiments to seek the optimal dengue vaccination ratio that could benefit the control of both diseases. Ultimately, we aim to provide vaccination guidelines for dengue prevalent regions to reduce local Zika cases.

Date: November 17, 2016
Time: 10:30 to 11:30 am
Location: CB 126

Basic Introduction of LDA Topic Model on Text Mining

Bowen Sun
Master of Science Candidate
Simon Fraser University, British Columbia, Canada.

Abstract: In machine learning and natural language processing, a topic model is a type of statistical model for discovering the abstract ‘topics’ that occur in a collection of documents. Intuitively, given a document is about some topic, the occurrence of some particular words is expected to be more or less frequent. Thus, the topic model is very useful in the content exploration or key point extraction from large document corpora. In this presentation, I will give an introduction of some basic process in text mining and the LDA topic model theory

Date: November 10, 2016
Time: 10:30 to 11:30 am
Location: CB 126

Modelling antibody enhancement and its consequence for integrated vector borne disease control

Biao Tang
Visiting PhD Candidate
Xi’an Jiaotong University, China,
York Univerisy and Fields Institution, Toronto, Canada

Abstract: We present our recent work on modelling co-infection of diseases sharing the same vector, and the consequence of antibody enhancement for integrated intervention including vaccination, using dengue and Zika as references. This is based on joint work with Yanni Xiao and Jianhong Wu.

Date: October 13, 2016
Time: 10:30 to 11:30 am
Location: CB 126

Mathematical modelling of cross-diffusion in biofilms

Kazi Rahman
Post-doctoral Fellow
York University and Ryerson University
Toronto, Canada

Abstract: We propose a deterministic continuum model for mixed culture biofilms where movement of one species is affected by the presence of the other. Two derivations of this new model are presented. One derivation is based on the continuous time, discrete space master equation and the other one is based on the equations of conservation of mass and momentum. Starting from both viewpoints, we derive the same dual-species diffusion-reaction model for biofilms that comprises three non-standard diffusion effects: (i) degeneracy as the local biomass density vanishes, (ii) a super-diffusion singularity as the local biomass density approaches its a priori known maximum, and (iii) non-linear cross-diffusion. (i) describes the finite speed of propagation of the biofilm/water interface, (ii) describes volume filling effects, and (iii) describes the mixing of both biomass species. We present a numerical method for this highly nonlinear PDE model of biofilm that can tackle these three nonlinear diffusion effects. To investigate the effect of the new model feature, we study the role of the cross-diffusion terms in numerical simulations of three biofilm models: competition, allelopathy, and a mixed system formed by anaerobic and an anaerobic species. In all three systems we observe that accounting for cross-diffusion affects local biofilm structure, in particular the relative local distribution of biomass, but it does not affect overall lumped quantities such as the total amount of biomass in the system. As an application, our highly nonlinear density dependent cross-diffusion model is used in order to incorporate an experimental observation in models of disinfection of microbial biofilms. An extended reaction kinetics based on carbon consumption during disinfection is introduced. Our simulations show that the extended model captures the experimental observation, and suggest that the consumption of carbon substrates during inactivation due to antibiotics helps biofilms to survive and re-grow. Finally, as an extension of dual-species model, a generalized cross-diffusion model of k interacting species is derived considering the continuous time and discrete space master equation passing to the continuous limit. Moreover, a criterion for preserving the positivity of the solution of this type of generalized cross-diffusion model is presented.

Date: October 6, 2016
Time: 10:30 to 11:30 am
Location: CB 126

Neural Dynamics and Optimization of Online Advertising Systems

Yong Yang
PhD Candidate
York University
Toronto, Canada

Abstract: Real time Bidding (RTB) is a relatively new advertising technology that allows online advertising to be purchased and served on the fly. RTB ingests and distributes impressions from thousands of parameters simultaneously. Infersystems is an algorithm provider whose technologies apply nonparametric statistics to improve the performance of both Demand Side Platforms (DSPs) and Supply Side Platforms (SSPs). InferSystems generates media buying and optimization rules to minimize the cost per action (CPA) and cost per click (CPC) of a digital advertising campaign. These results are accomplished by using InferSystems proprietary, predictive analytic and decision engine (i.e., Infer Engine) that is able to predict super rare events from sparse data. Based on a variety of data training techniques, the Infer Engine automatically outputs a decision table of media buying rules. Buying rules are the number of combinations of parameters, such as country, banner code, banner position, landing url and bidding time. When users access the web page, all user information is uploaded to the web server. Infer Engine attempts to match impressions and the rules in the decision table. These generated media buying rules are able to predict which impressions are the most likely ones to be the clicked advertisements. As time goes by, these buying rules will fail to work and hence we expect to see a dramatic drop in the number of impressions if these buying rules are not revised / replaced. The objective of this thesis is two folds: to predict click impression dynamics, and to quantify the effectiveness of media buying rules and the decay of these rules. The model proposed is similar to, but different from, the classical epidemiological models for infectious diseases, this is due to the similarity between the considered click impression dynamics and the disease infection exposure dynamics.

Date: September 22, 2016
Time: 10:30 to 11:30 am
Location: CB 126

LIAM Initiatives during 2016-2017

Dr. Jianhong Wu
University Distinguished Research Professor
Canada Research Chair in Industrial and Applied Mathematics
Laboratory for Industrial and Applied Mathematics (LIAM)
Department of Mathematics and Statistics
York University, Toronto, Canada

Date: September 16, 2016
Time: 11:30 am to 12:30 pm
Location: CB 126

Neural Network Based Approach for Subspace Clustering of High Dimensional Data

Bowen Sun
Visiting M.Sc. Student
Simon Fraser University
Vancouver, Canada

Abstract: We are living in a data/information driven society, and every aspect of our life greatly depends on our ability to collect, analyze and understand large sets of data and information. These large data sets with high dimensions arise naturally from a variety of fields, such as bioinformatics, text mining. And the traditional clustering algorithms can’t work effectively for these types of data because of the well- known problem, the curse of dimensionality. In this talk, I am going to introduce a neural network architecture, PART developed by Professor Wu and his LIAM team, which is based on the ART developed by Carpenter and Grossberg. I will also discuss the algorithm for clustering high dimensional data sets.

Date: April 22, 2016
Time: 11:30 am to 12:30 pm
Location: CB 126

Mathematical Solutions of Pharmacokinetic Models When Nonlinearity Is An Issue

Xiaotian Wu
Visiting Professor
University of Montreal
Montreal, Canada

Abstract: Analytical solutions of pharmacokinetic models are appealing since they provide a clear and direct way to reveal the relationship between different model components, and greatly improve the process of drug development and drug design. In this talk, I will present mathematical solutions, including time-course of drug concentration and estimation of key pharmacokinetics parameters, of some pharmacokinetic models where nonlinear elimination is an important factor for some drug disposition, typically hormone drugs such as granulocyte colony-stimulating factor (G-CSF). This is a joint work with Professors Fahima Nekka and Jun Li at Université de Montréal.

Date: April 8, 2016
Time: 11:30 am to 12:30 pm
Location: CB 126

Predictive Modelling of Health Clusters for Chronic Disease Management

Yawen Xu
Postdoctoral fellow
York University and Manifolds Data Mining Inc.
Toronto, Canada

Abstract: Join Yawen Xu, York University, and Ted Hains & Zhen Mei, Manifold Data Mining Inc. lecture on health clusters of patients with chronic diseases, particularly diabetes and heart diseases. They will introduce a predictive modelling technique for classifying a patient into clusters, based on their demographics, mental health and stress, health outcomes, social connection, motivation and lifestyles.

Date: April 1, 2016
Time: 2:30 pm to 4:30 pm
Location: York Lanes Room 280N

Mathematical Modeling of Infectious Diseases From the Pre-Vaccine to Vaccine Era

Felicia Maria G. Magpantay
Assistant Professor
University of Manitoba
Manitoba, Canada

Abstract: The dynamics of vaccine-preventable diseases depend on the underlying disease process and the nature of the vaccine. In this talk I will discuss imperfect vaccines and the epidemiological consequences of different modes of vaccine failure. In particular, I will focus on the dynamics during the transition from the pre-vaccine to vaccine era and some new methodologies for dealing with incomplete data during this period.

I will also present an application to pertussis, a childhood disease that was once considered a candidate for eradication. This highly infectious disease is still a significant cause of child mortality in the world, and has been reemerging in some countries that maintain high vaccination coverage (e.g. USA, UK). Recent events have highlighted how much we still do not know about the mechanics of this disease and the type of immunity rendered by infection and vaccination. I will discuss some of the progress we have made in fitting a general stochastic model of pertussis, and the ideas behind the likelihood-based statistical inference methods (trajectory matching and iterated filtering) used to estimate the vaccine parameters.

Date: March 28, 2016
Time: 2:00 pm to 3:00 pm
Location: CB 126

Self-Excited Vibrations in Damped Wave Equations

Nemanja Kosovalic
Assistant Professor
University of South Alabama
Alabama, United States

Abstract: Over the last fifty years much work has been devoted to the study of forced vibrations in damped wave equations. From the mechanical point of view, external forcing is the simplest way of putting energy back into the system to balance out the friction, which results in a global time periodic solution whose amplitude does not decay. Another way of putting energy back into the system results from the presence of a restoring force with time delay. This leads to ‘self-excited’ vibrations. We discuss some aspects of self-excited vibrations for damped wave equations.

Date: March 18, 2016
Time: 11:30 am to 12:30 pm
Location: CB 126

The Evolution of Antimicrobial De-escalation: Part II

Lindsey Falk
Masters of Public Health (Epidemiology) Candidate
University of Toronto
Toronto, Canada

Abstract: Antibiotic resistance is a prominent issue in healthcare and there is a need to identify strategies that reduce resistance to broad spectrum antibiotics without compromising patient outcomes. This talk builds on a previous seminar given by Xi Huo and Josie Hughes, where a transmission model of P. aeruginosa in an intensive care unit (ICU) was presented to explore the evolutionary and ecologic impacts of antimicrobial de-escalation. De-escalation is a treatment strategy that aims to preserve the efficacy of broad spectrum antibiotics by switching patients to narrower agents. Although it has been applied widely in ICUs, the impacts are poorly understood. I will present the results from the model, which compares the de-escalation strategy to usual care under two different de-escalation approaches. As the mathematical model has previously been discussed, the focus will be on the clinical and public health implications of the model.

Date: March 11, 2016
Time: 11:30 am to 12:30 pm
Location: CB 126

Mathematical Modeling of With-In Host Dynamics of Listeria monocytogenes

Ashrafur Rahman
Postdoctoral Fellow
York University
Toronto, Canada

Abstract: Listeriosis is a potential food-borne disease caused by L. monocytogenes. The disease is an important public health problem as it poses a severe risk to certain populations including pregnant women, older adults, and individuals with a weakened immune systems. An individual can be infected with L. monocytogenes after consuming contaminated food. The bacteria can colonize in the intestines and reach the liver, spleen and placenta via the blood and lymphatic vessels. In this talk, I will outline modeling approaches of the Listeria invasion into the gut and its translocation at different organs. I will highlight the role of the immune responses and the inoculation doses that determine the difference of infection, and characterize the critical transition of listeriosis from mild to severe. This is based on joint work with D. Munther, J. Wu and a team of scientists from the Public Health Agency of Canada.

Date: March 4, 2016
Time: 11:30 am to 12:30 pm
Location: CB 126

Existence and Uniqueness of Mild and Strict Solutions for Abstract Differential Equations with State Dependent Delay

Michelle Pierri
Professor
São Paulo University
São Paulo, Brazil

Abstract: The theory of differential equations with delay is one of many important branches of the theory of differential questions. Recently, a new class of delay equations with a state-dependent delay (SDD) has attracted much attention of researchers. The study of ordinary and partial differential equations with state dependent delay differ from the case of ordinary and partial differential equations with constant or time-dependent delays. In this talk, we present some results related the existence of mild and strict solutions for abstract differential equations with state dependent delay. Our main result is concerning the existence of alpha-Holder strict solutions for this class of problems.

Date: February 26, 2016
Time: 11:30 am to 12:30 pm
Location: CB 126

Population Dynamics of Borrelia burgdorferi in Lyme disease

Jemisa Sadiku
PhD Candidate
Department of Mathematics and Statistics, York University
Toronto, Canada

Abstract: Many chronic inflammatory diseases are known to be caused by persistent bacterial or viral infections. A well-studied example is the tick-borne infection by the gram-negative spirochaetes of the genus Borrelia in humans and other mammals. It causes severe symptoms of chronic inflammation and subsequent tissue damage (Lyme Disease), particularly in large joints and the central nervous system, but also in the heart and other tissues of untreated patients. Although killed efficiently by human phagocytic cells in vitro, Borrelia exhibits a remarkably high infectivity in mice and men. In experimentally infected mice, the first immune response almost clears the infection. However, approximately 1 week post infection, the bacterial population recovers and reaches an even larger size before entering the chronic phase. We discussed a mathematical model (Binder et al., Frontiers in Microbiology, 2012) that describes the bacterial growth and the immune response against Borrelia burgdorferi in the C3H mouse strain that has been established as an experimental model for Lyme disease. The peculiar dynamics of the infection excludes two possible mechanistic explanations for the regrowth of the almost cleared bacteria. Neither the hypothesis of bacterial dissemination to different tissues nor a limitation of phagocytic capacity were compatible with experiment. The mathematical model predicts that Borrelia recovers from the strong initial immune response by the regrowth of an immune-resistant sub-population of the bacteria. The chronic phase appears as an equilibration of bacterial growth and adaptive immunity. This result has major implications for the development of the chronic phase of Borrelia infections as well as on potential protective clinical interventions.

Date: February 19, 2016
Time: 11:30 am to 12:30 pm
Location: CB 126

Research on Key Technologues to Quantify, Simulate, and Standardize Acupuncture Manipulations

Yine He
Faculty
Shanghai University of Traditional Chinese Medicine
Shanghai, China

Abstract: In this talk, I will introduce the basic concepts of acupuncture and how it can be applied to heal some diseases in China. I will talk about our project and the problems we encountered during clinical activities and research work. The project was based on previous studies on acupuncture techniques and we wanted to improve the hardware devices for parameter acquisition. We improved the software by quantifying parameters, taking videos, and collecting commentary of expert acupuncturists. Afterwards, we established an integrated database of acupuncture specialists and their manipulations. To this database, we applied data mining technologies to analyze the data, extracted common characteristics, and established a mathematical model which describes the features of good acupuncture manipulations. Finally, we applied simulation technologies to have established an experimental research platform based on the mathematical model. The platform can be used to collect and analyze data, and to simulate excellent acupuncture manipulations.

Date: February 5, 2016
Time: 11:30 am to 12:30 pm
Location: CB 126

Modelling the Post-Treatment Control of HIV Infection Using a Within-Host Model with Latent Reservoir and Immune Impairment

Shaoli Wang
Lecturer
School of Mathematics and Statistics, Henan University
Henan, China

Abstract: I will give a simplified within-host model with latent reservoir and immune impairment to explain the post-treatment immune control by exploring the bistability of the model. Mathematically, we show if the basic infection reproductive number, R_{0}, is less than one, the uninfected equilibrium of the proposed model is globally asymptotically stable, which means that the virus will die out. If R_{0} is greater than one, we can obtain two additional thresholds: the post-treatment immune control threshold and the elite control threshold. If the proliferation rate of CTLs is less than the post-treatment immune control threshold, the positive equilibria does not exist, the immune free equilibrium is stable, and the system will have virus rebound. If the proliferation rate of CTLs is within the bistable interval (between the two additional thresholds) and the initial virus population is low, then the system will be under post-treatment immune control. While the proliferation rate of CTLs is greater than the elite control threshold, the positive immune equilibrium is stable, the immune free equilibrium is unstable, and the system will be under elite control.

Date: January 29, 2016
Time: 11:30 am to 12:30 pm
Location: CB 126

Existence and Uniqueness of Solutions for Abstract Neutral Equations with State Dependent Delay

Eduardo Hernández
Professor
São Paulo University
São Paulo, Brazil

Abstract: In this seminar we present some results on the existence and uniqueness of strict solutions for a class of abstract neutral equations with state dependent delay with applications to partial neutral functional differential equations.

Date: January 22, 2016
Time: 11:30 am to 12:30 pm
Location: CB 126

Local Bifurcation Theory for Some Nonreversible Wave Equations

Nemanja Kosovalic
Assistant Professor
University of South Alabama
Alabama, United States

Abstract: Over the last fifty years a huge effort has been devoted to the study of the local bifurcation of periodic and quasi-periodic
solutions for reversible wave equations. Despite this effort, there are gaps in what is currently known about the nonreversible counterpart. Nonreversible wave equations generally include wave equations having either time delay or damping terms. We discuss some results in this direction and some open problems. This work presented is collaborative work with Dr. Brian Pigott and Dr. Chris Lin.

Date: December 21, 2015
Time: 11:30 am to 12:30 pm
Location: CB 126

The Evolutionary Ecology of Antimicrobial Descalation

Josie Hughes
Post-doctoral Fellow
Mount Sinai Hospital

Xi Huo
Post-doctoral Fellow
York University and Ryerson University

Abstract: We model the transmission of P. aeruginosa in intensive care units (ICUs) with deescalation as the major antibiotic treatment strategy. That is, empirical therapy is initiated when a patient is infected with P. aeruginosa, right after the laboratory test results become available, the definitive therapy will be de-escalated – the broad-spectrum antibiotic for empirical therapy is switched to a narrow-spectrum antibiotic if possible. De-escalation is a treatment strategy that have been applied widely in ICUs, with the aim of reducing the risk of super-infection and preserve the efficacy of broad spectrum drugs. It has been considered as a potential way of reducing antibiotic use and antimicrobial resistance in ICUs. This is a project of the Development of an Antimicrobial Resistance Diversity Index (ARDI) led by Prof. Jianhong Wu.

Date: December 11, 2015
Time: 11:30 am to 12:30 pm
Location: CB 126

Flocking, Flocking Bifurcation and Flocking Switches in a Two-Agent Flock with Processing Delay

Xiao Wang
Professor
College of Science, National University for Defense Technology
Changsha, China

Abstract: Necessary and sufficient conditions are established for a two-agent flock model with processing delay to admit a time-asymptotic flocking. The results provide a relation based on which proper initial positions and velocities can be selected to form a flocking with predetermined position displacement distance. It is shown that the processing delay can terminate a flocking, can induce a flocking and can lead to a flocking bifurcation resulting a periodic flocking. It is also shown that the processing delay can induce flocking switches in the sense that as the processing delay varies, the flocking may follow a switching pattern as no flocking-flocking-periodic flocking-flocking-divergence.

Date: November 27, 2015
Time: 11:30 am to 12:30 pm
Location: CB 126

Multiple-Platform Data Integration Method with Applications to Combined Analysis of Microarray and Proteomic Data

Yawen Xu
Postdoctoral Fellow
Department of Mathematics and Statistics, York University
Toronto, Canada

Abstract: It’s desirable in genomic studies to select biomarkers that differentiate between normal and diseased populations based on related data sets from different platforms. Most recently developed integration methods focus on correlation analyses between gene and protein expression profiles. These methods select biomarkers with concordant behavior but do not directly select differentially expressed biomarkers. Other methods combine statistical evidence in terms of ranks and p-values, but they don’t account for the dependency relationships among the data across platforms. We propose an integration method to perform hypothesis testing and biomarkers selection based on multi-platform data sets observed from normal and diseased populations.

Date: November 13, 2015
Time: 11:30 am to 12:30 pm
Location: CB 126

Stability or Instability of Switched Systems with Time-Delay Using Fast and Random Switches

Yao Guo
Postdoctoral Fellow
Department of Mathematics and Statistics
Toronto, Canada

Abstract: This talk first presents examples that systems even with time delays switching among stable subsystems can be destabilized by fast switches. To illustrate this phenomenon, we introduce a prototype model which includes a linear random switched system with time-delay, and theoretically establish conditions under which the switched system can still be either unstable or stable by using certain sets of switches. Standard tools of stochastic theory are utilized in theoretical arguments, including Doob’s Optimal Stopping Theorem. In addition, we explain intuitively how this phenomenon happens and provide numerical simulations to reinforce our theoretical results.

Date: October 9, 2015
Time: 11:30 am to 12:30 pm
Location: Ross N638

Bistability in Ideological Conflict

Shaoli Wang
Henan University
Henan, China

Abstract: In this talk, we analyze the dynamics of the ideological model provided by Marvel et al. [1]. We show that bistability appears when the constant fraction of zealots, p, is less than a critical value p=pc. In this case, when a subpopulation crosses certain boundaries, its stable equilibrium switches accordingly. We also prove that in the case of p=pc, a saddle-node bifurcation replaces the bistability, and leaves a unique stable equilibrium, which means the entire population reaches a consensus. Simulations show that a system with two zealots, p1 and p2, is bistable with strictly positive equilibrium.

Date: September 30, 2015
Time: 2:30 pm to 3:30 pm
Location: TEL 5021A

DISTINGUISHED LECTURE: Traveling Waves in Isothermal Diffusion Systems: Existence, Stability and Oscillations

Yuanwei Qi
University of Central Florida
Florida, USA

Abstract: In this talk I shall present some of the most recent results my and collaborators and I have proved in the last a few years. In particular, we show a promising model proposed by a leading world authority in chemical engineering, Prof. Gary of U. Cambridge, FRS, has very rich structures and the analytic study proves to be far more challenge than the old model.

Date: September 30, 2015
Time: 4:00 pm to 4:30 pm
Location: TEL 5021A

Some Points on Traveling Wave Front of Reaction-Diffusion Systems with Delay: The Threshold Dynamics

Ruili Feng
PhD candidate, University of Science and Technology of China

Abstract:

Date: September 18, 2015
Time: 11:30 am to 12:00 pm
Location: Ross N638

Building an Assessment Model for a Movie Quality Rating

Yong Yang

Abstract: SINA weibo (新浪微博)and Film and state administration of press, publication, radio, film and Television of The people’s republic of china(国家广电总局) propose a project on movie quality rating assessment. SINA Weibo is one of the most popular Chinese microblogging website in china. SINA team wants to build a model for the movie quality rating .Based on this project, I will introduce a dataset and propose a method. Any suggestions or methods are welcome.

Date: September 18, 2015
Time: 12:00 pm to 12:30 pm
Location: Ross N638

Consensus and Clustering of Linear Leader-Following System with Delay

Yicheng Liu
Associate Professor
College of Science, National University of Defense Technology
Changsha, China

Abstract: Two or more groups with different initial opinion values would develop into a common consensus if there are special communications and interactions between groups. What adjacent structures within groups make them reaching a consensus is an important issue. In this talk, we present an opinion consensus and clustering problem between two groups, say leader group and following group. The evolution of opinions is described as a normalized continuous dynamical system involving a distributed delay. With hypothesis that the normal Laplacian matrix has a semisimple zero eigenvalue, we find that the coupled system reaches an unconditional consensus if and only if the multiplicity of zero is 1, and reaches a conditional consensus if the coupled structure is multipartite. Also, we will mention the relationship between consensus value of coupled group and consensus values of each group. An analytic consensus value formula is deduced by using the eigenvector analysis method. In results, we find the consensus value falls into the interval of leader’s and following’s consensus values. Meanwhile, the influence strength between groups has sensitively affected consensus value of the coupled system.

Date: September 11, 2015
Time: 11:30 am to 12:30 pm
Location: Ross N638