Oral Sessions

Session 1: 

Assistant Prof Yinghao Pan 

Postdoctoral Researcher Paula Oliveira

Assistant Professor Xiang Chen

Session 2:

Assistant Prof Vlad Margarint

Assistant Professor Sebastien Bossu

Associate Professor Taner Tunc

We have proposed an unprecedented deterministic model of Lassa Hemorrhagic fever (LHF) model with nonlinear force of LHF infection to capture the transmission dynamics and long-term effects of the disease. The Qualitative analyses we have conveyed on this model using well-established methods viz: Cauchy's differential theorem, Birkhoff & Rota's theorems verify and reveal the well-posedness of the model respectively. We established that an LHF-free equilibrium termed the disease-free equilibrium (DFE) exists for this model and this equilibrium, however, from our stability analyses, tends to be stable when the basic reproduction number computed via the next generation matrix method is less than unity (one); and unstable otherwise. Furthermore, we have carried out a sensitivity analysis to check for the variation effects of the model parameters when increased or decreased using the normalized forward-sensitivity index; unraveling the most sensitive parameters which requires the attention of the healthcare workers as; the effective contact rates and the rodents’ recruitment rate. After which numerical simulations of the model were carried out to verify our qualitative analyses (Stability and sensitivity analysis) and to study the dynamical behavior of the model; showing that the presence of saturation instantaneously causes the system to approach a DFE/LHF-Free equilibrium. From these qualitative analyses and numerical simulation results, we recommend early intervention and early treatment of Lassa hemorrhagic virus infection (LAHV) with Ribavirin on the infected, maximum hygiene practices and periodic evacuation of rodents in households in order to curb the recruitment of wild/rodents. 

We have proposed an unprecedented deterministic model of Lassa Hemorrhagic fever (LHF) model with nonlinear force of LHF infection to capture the transmission dynamics and long-term effects of the disease. The Qualitative analyses we have conveyed on this model using well-established methods viz: Cauchy's differential theorem, Birkhoff & Rota's theorems verify and reveal the well-posedness of the model respectively. We established that an LHF-free equilibrium termed the disease-free equilibrium (DFE) exists for this model and this equilibrium, however, from our stability analyses, tends to be stable when the basic reproduction number computed via the next generation matrix method is less than unity (one); and unstable otherwise. Furthermore, we have carried out a sensitivity analysis to check for the variation effects of the model parameters when increased or decreased using the normalized forward-sensitivity index; unraveling the most sensitive parameters which requires the attention of the healthcare workers as; the effective contact rates and the rodents’ recruitment rate. After which numerical simulations of the model were carried out to verify our qualitative analyses (Stability and sensitivity analysis) and to study the dynamical behavior of the model; showing that the presence of saturation instantaneously causes the system to approach a DFE/LHF-Free equilibrium. From these qualitative analyses and numerical simulation results, we recommend early intervention and early treatment of Lassa hemorrhagic virus infection (LAHV) with Ribavirin on the infected, maximum hygiene practices and periodic evacuation of rodents in households in order to curb the recruitment of wild/rodents. 

Realistic human movement, essential for applications like video games, computer graphics, and robotic simulations, is advanced through 3D human pose estimation, which measures joint angles and body segment relationships. This review evaluates deep learning-based 3D Human Pose Estimation methods, highlighting their strengths, weaknesses, and applications in fields like surveillance and healthcare. It concludes by discussing ongoing challenges and future research directions to enhance the accuracy and performance of these systems. 

PCR can fail because of poor primer design. Nearest-neighbor thermodynamic properties, picking conserved regions, and filtration via penalty of oligonucleotides form the basis for 4 primer design. DeGenPrime-Ez is a user-friendly graphical user interface that creates high quality PCR primer design based on DeGenPrime [1]. Our interface can utilize multiple sequence alignment formats and direct sequences, expanding the target range for a single primer set. Primer design and refinement is based on thermodynamic properties, filtration metrics, penalty scoring against degenerate bases, and conserved region finding of any proposed primer. It has filters for degeneracy, repeated k-mers, relative GC content, and temperature. Minimal penalty scoring is included according to secondary structure self-dimerization metrics, GC clamping, tri- and tetra-loop hairpins, and internal repetition. DeGenPrime-Ez is written with Tkinter and CustomTkinter APIs with accessibility in mind. Our program provides customizable interface formatting via pre-packaged JSON files which promotes accessibility for neurodivergent and/or colorblind individuals. The GUI is easily accessible to scientists without a computational background, and output is customizable. DeGenPrime-Ez unlocks the tree of life with a level of accessibility not previously seen in other tools.  

Functional data has received significant attention due to its frequent occurrence in modern applications, such as functional magnetic resonance imaging (fMRI). Existing research in this area has primarily focused on traditional methods, with a surprising lack of investigation into quantile regression. The infinite-dimensional nature of functional data necessitates the use of dimension reduction techniques. However, the classical sufficient dimension reduction techniques for functional data are typically restricted to quantitative predictors and do not incorporate categorical predictors. To overcome this limitation, we extend the partial quantile dimension reduction techniques to functional cases, enabling the analysis of data involving both quantitative and categorical predictor variables. Specifically, we introduce the concept of the τth functional central partial quantile subspace (τ-FCPQS) and propose a two-step algorithm for its estimation. First, we apply an established functional quantile dimension reduction method to estimate the τth functional central quantile subspace for each subgroup and then, we perform an eigenfunction problem. Moreover, we derive the convergence rates of the proposed estimators and demonstrate their finite sample performances through simulation studies and real data set based on fMRI. Results suggest that (1) the efficiency of the proposed methodology increases with the sample size n and decreases with the number of predictors p; (2) the proposed method has 4 finite sample performance and provides a strong alternative to functional data analysis; (3) it is comparable to existing methods. This study addresses the gap in dimension reduction techniques for the conditional quantiles of functional data, particularly in the context of mixed predictor types, while also presenting a methodologically sound and statistically significant advancement in the analysis of such data.

Decentralized systems play a significant role in the public and private sectors that serve different needs of organizations. The heart of those systems in any system is the dissemination protocol. Hyperledger Fabric is one of the most popular production-ready distribution network systems. Though Hyperledger Fabric is intended to have any consensus protocol as a pluggable module, it does not provide technical guidance on how to plug a consensus module there. However, Fabric used Kafka and Raft [1] for the Ordering service consensus protocol. Both are crash fault tolerant (CFT) protocols, especially Raft, which is based on Paxos [2] and does not support Byzantine fault problems [3]. We developed a new fabric system that successfully integrates the Byzantine Fault Tolerant consensus protocol using BDLS [4] into Hyperledger Fabrics. The consensus protocol BDLS is among the most efficient and promising BFT protocols for blockchains. We compare the performances of Raft and BDLS in Hyperledger Fabric system. The study shows that Hyperledger Fabric, employing BDLS fault tolerance protocols, shows outperformance results for a BFT module. Matches the performance of the existing Raft-based Fabric(fig.2). Our BDLS Fabric implementation, featuring BFT consensus, achieves a throughput of 1135 transactions per second (TPS), nearly on par with Raft-Fabric's 1158 TPS (fig.1), despite Raft-Fabric lacking support for BFT consensus. The evaluation of a single node performance of BDLS and RAFT (fig.3) shows outperformance results, which bring the secured BFT module closer to the fast CFT module than ever before.