Past Student Presentations
Date: Thursday, October 30, 2025
Time: 10:00 am – 12:00 pm
Room: Online via Google Meet: meet.google.com/hhz-doei-gpx
Number of Presentations: 4
Presentation Details – not listed in any particular order
Presentation 1:
Title of Presentation: Over-Squashing in GNNs and Causal Inference of Rewiring Strategies
Presenter Name: Danial Saber (MSc Candidate, Computer Science, Ontario Tech University)
Supervisor: Dr. Amirali Salehi-Abari (Faculty of Business and Information Technology, Ontario Tech University)
Abstract: Graph neural networks (GNNs) have exhibited state-of-the-art performance across wide-range of domains such as recommender systems, material design, and drug repurposing. Yet message-passing GNNs suffer from over-squashing—exponential compression of long-range information from distant nodes—which limits expressivity. Rewiring techniques can ease this bottleneck; but their practical impacts are unclear due to the lack of a direct empirical oversquashing metric. We propose a rigorous, topology-focused method for assessing over-squashing between node pairs using the decay rate of their mutual sensitivity. We then extend these pairwise assessments to four graph-level statistics (prevalence, intensity, variability, extremity). Coupling these metrics with a within-graph causal design, we quantify how rewiring strategies affect oversquashing on diverse graph- and node-classification benchmarks. Our extensive empirical analyses show that most graph classification datasets suffer from over-squashing (but to various extents), and rewiring effectively mitigates it—though the degree of mitigation, and its translation into performance gains, varies by dataset and method. We also found that over-squashing is less notable in node classification datasets, where rewiring often increases oversquashing, and performance variations are uncorrelated with oversquashing changes. These findings suggest that rewiring is most beneficial when over-squashing is both substantial and corrected with restraint—while overly aggressive rewiring, or rewiring applied to minimally over-squashed graphs, is unlikely to help and may even harm performance. Our plug-and-play diagnostic tool lets practitioners decide—before any training—whether rewiring is likely to pay off.
Presentation 2:
Title of Presentation: Examining the Haptic Prototyping Process using a Novel Open Source Haptic Creation Tool
Presenter Name: Sam Canonaco (MSc Candidate, Computer Science, Ontario Tech University)
Supervisor: Dr. Pejman Mirza-Babaei (Faculty of Business and Information Technology, Ontario Tech University)
Abstract: Haptics have advanced rapidly in the past decade. They are becoming an essential way of delivering tactile information on our phones and game controllers. Haptic devices can be as small as the vibrate feature on a smartwatch or as large and specialized as theatre seats that move and rumble to add more immersion to media experiences. Advanced haptic devices are more prevalent than ever and AAA game developers have been embracing them more and more each year. Today, consumers and developers can purchase an advanced haptic device at nearly any electronics store for a relatively low cost. Our research aims to examine how haptic effects can be prototyped by student developers, with a broad goal of making haptics more accessible to aspiring and independent game developers. Haptic effect creation is an upcoming complex area of research, as most people do not have experience discussing and using haptics of their own. We created a software tool called SIGHTT, which enables the testing of haptic effects on a standard DualSense gamepad. This tool allowed our participants to create haptic effects for a variety of different scenarios that commonly occur in games, despite most of them having no experience creating effects previously. This, among with other observations and patterns, demonstrated the potential of haptics if developers simply have access to the right tools. Increased developer use of haptics has the potential to advance the way that haptics are discussed and help develop knowledge in the field. We hope and believe that this project can empower more developers to make better use of haptic devices in the future and bring a new layer of perception to our digital experiences.
Presentation 3:
Title of Presentation: Network Disaster Recovery via Dynamic Service-Aware Risk Management
Presenter Name: Sara Taghavi Motlagh (PhD Candidate, Computer Science, Ontario Tech University)
Supervisor: Dr. Shahram Heydari (Faculty of Business and Information Technology, Ontario Tech University)
Abstract: Communication networks are increasingly exposed to natural and human-induced disasters, which result in service disruption and economic losses. Most existing disaster recovery frameworks treat risk as a static factor, overlooking the fact that hazard conditions evolve over time. For example, in earthquakes, failures propagate outward from the epicenter across successive time steps, creating time-varying and correlated risk for network links. In practice, routing must therefore be both dynamic and service-aware, reflecting class priorities, SLA targets, and capacity constraints, yet existing models rarely integrate these elements in a unified optimization framework. This thesis formulates dynamic, service-aware routing under disaster conditions as a multi-objective mixed-integer nonlinear program (MINLP). The model jointly optimizes five objectives: maximizing revenue and survivability, while minimizing service-aware risk, overutilization cost, and SLA penalties, subject to capacity and service-level constraints. A dynamic risk model is introduced that considers time-varying link failure probabilities and is weighted by service priorities. This includes both hazard progress and the importance of impacted services. SLA compliance is preserved by class of service availability constraints, which guarantee the preservation of high-priority traffic during disruptions. We solve the model using the Non-dominated Sorting Genetic Algorithm II (NSGA-II), which approximates the Pareto front and provides a decision-support layer for service providers. Experiments are conducted on real-world network topologies (ERnet, France, Giul39) under staged and dynamic failure scenarios. The results show that adaptive routing shifts flows from high-risk to safer links as hazards evolve, improving survivability and reducing SLA penalties while explicitly exposing trade-offs with revenue and overutilization. Sensitivity analyses show how link capacity, service request volume, and failure probability affect the balance among the five objectives. Dynamic risk, service differentiation, and capacity constraints are embedded in a single multi-objective formulation, enabling decision-makers to explore balanced recovery strategies rather than committing to a single prescriptive solution. The approach provides a practical tool for evaluating disaster-resilient routing policies under uncertainty, which helps strengthen communication networks against evolving natural and human-induced disruptions.
Presentation 4:
Title of Presentation: How is fairness measured in practice? A survey and critique of fairness analyses applied to predictive algorithms
Presenter Name: Andrew Putman (PhD Candidate, Computer Science, Ontario Tech University)
Supervisor: Dr. Peter Lewis(Faculty of Business and Information Technology, Ontario Tech University), and Dr. David Rudoler (Faculty of Health Sciences, Ontario Tech University)
Abstract: The increasingly rapid pace of research and development of prediction algorithms over the past decade has spurred the adoption of such predictive algorithms across seemingly every industry and aspect of life. This growth and increase of public awareness have also brought into focus concerns around the potential bias and fairness impacts of these algorithms. This cross-disciplinary survey assesses peer-reviewed publications from the last 10 years which have implemented a form of bias and fairness analysis on a predictive algorithm using real world data. We begin with a descriptive overview of the fields of research, types of publication, and the types of outcomes being predicted. We then engage in a more detailed exploration of the fairness analyses that were applied, the metrics used, the evidence provided to justify these selections. We also discuss the contexts in which these analyses were conducted. We conclude by highlighting the stark lack of inclusion of perspectives from people who are impacted by the algorithms being analyzed and discuss directions for future work in this area.
Date: Tuesday, January 28, 2025
Time: 2:00 pm – 3:30 pm
Room: Science Board Room (UA 4170)
Number of Presentations: 3
Presentation 1:
Title of Presentation: Novel optimizers for multi-objective machine learning
Presenter Name: Farzaneh Nikbakhtsarvestani (PhD Candidate, Computer Science, Ontario Tech University)
Supervisor: Dr. Mehran Ebrahimi (Faculty of Science, Ontario Tech University) and Dr. Shahryar Rahnamayn (Faculty of Engineering and Applied Science, Ontario Tech University)
Abstract: Multi-objective optimization is a prevalent challenge in the area of deep learning. There is a lack of robust multi-objective optimization methods applicable in deep learning capable of training networks by simultaneously optimizing conflicting multiple loss functions. Its applications include a wide range of deep neural network branches such as multiloss, multi-task, multi-modal, and crossmodal learning. In this thesis, we develop two novel optimizers: Multi-objective Adam Optimizer (MAdam) and Opposition-base Multi-objective Adam Optimizer (OMAdam). MAdam is a multi-objective extension of the well-known Adam optimization algorithm. MAdam is a classical population-based approach that uses the gradient information of multiple objectives to accelerate population convergence toward an optimal minimum. The method applied a non-dominated sorting algorithm to keep selective population and improve the diversity across the landscape. In OMAdam, opposition based scheme was introduced into MAdam framework as global search is necessary for escaping local optima in gradient-based multi-objective optimization approaches. The Opposition-based MAdam, explores multiple directions over the landscape, that leads to independence from specific initialization. These two optimizers are scalable in every deep learning tasks. In a series of experiments for MAdam and OMAdam applications, we investigate multi objective binary and multi-class classification tasks using novel Madam and OMAdam optimizers. The aim of the study is to provide a range of solutions on the Pareto front to balance trade-offs between conflicting loss functions and to select corresponding model parameters that reflect the clinical priorities of practitioners (classes and sub-classes). In this regard, we employed these two novel optimizers as AI diagnostic techniques based on the 10-class MNIST and 2-class, 8-class, and 11-class MedMNIST datasets. In each experiment, the Pareto front of model parameters allows for variations in performance metrics and enable the prioritization of specific classes and sub-classes targeted by the end-users. MAdam and OMAdam optimizers achieve F1-scores comparable to those of the single objective Adam optimizer among their many solutions on the Pareto front. Also, we investigate the functionality of Madam and OMAdam optimizers in another deep learning task: Multi-objective regression task. In this study, we modify the optimization process to include conflicts losses for fitting error, model complexity, and Huber loss, which represent the model’s performance on data consistency, regularization, and resistance to outliers, respectively. we incorporate these conflicting loss terms to MAdam and OMAdam framework. These schemes, address the challenges of noise sensitivity and overfitting in scenarios dominated by collinearity or high-dimensional feature spaces. We demonstrate how these two novel optimizers can efficiently train artificial neural networks by balancing these conflicting objectives without the need for multiple retraining cycles. Our approach constructs a Pareto front that provides a spectrum of solutions, each offering a different trade-off between model’s accuracy , complexity and robustness. These methods broadens the applicability of MAdam and OMAdam optimizers in deep learning, particularly for critical applications like financial forecasting and healthcare predictions, where decision-making relies heavily on the robustness and accuracy of predictive models.
Presentation 2:
Title of Presentation: Balancing Security and Usability: A PII-Driven Framework for Password and Honeyword Generation in Modern Authentication Systems
Presenter Name: Hannah Adjei (PhD Candidate, Computer Science, Ontario Tech University
Supervisor: Dr. Miguel Vargas Martin
Abstract: The growing demand for strong yet easy-to-use authentication systems is pushing forward the development of innovative password-generation techniques to respond to the increasing threat of cyber-attacks and the critical need for user-friendly security measures.
This research investigates a novel approach to creating strong yet memorable and usable passwords that utilize Personally Identifiable Information (PII) extracted from user data. A systematic methodology was employed to evaluate PII-based passwords alongside randomly generated PII passwords (RaPII) and purely random passwords (Ra-Pass). Password strength was assessed using zxcvbn scores, cosine similarity measured honeywords’ closeness to actual passwords, and Levenshtein distance quantified memorability. Our comprehensive analysis, including ANOVA followed by Tukey HSD tests, underscored significant distinctions among the different password categories. Building on the results, which highlight the enhanced usability and robustness of passwords from PII, a comprehensive user study utilizing the System Usability Scale (SUS) provided more profound insights into users’ interactions and perceptions of these passwords. This research contributes a solid framework for developing PII-driven password and honeyword systems, integrating statistical analysis with a focus on user experience. We address current cybersecurity challenges and aim to set the foundation for future improvements in authentication technologies. By effectively blending the familiarity of PII with rigorous security protocols, it seeks to provide secure, user-focused solutions.
Presentation 3:
Title of Presentation: Gazed-enabled Implicit Interaction for Scatterplots
Presenter Name: Tania Shimabukuro (MSc Candidate, Computer Science, Ontario Tech University)
Supervisor: Dr. Christopher Collins (Faculty of Science, Ontario Tech University)
Abstract: How we look at the world can tell much about our focus and interests. Yet, this information is not used when displaying information visualizations. This research investigates novel display and interaction designs for create gaze-responsive visualizations optimized to support analysis of the items of interest. Specifically, we investigate four opportunities for implicit gaze-driven interaction for subtle visualization optimizations using eye tracking to improve standard scatter plot tasks: rendering order changes, hover speed modifications, local reference axes, and click accuracy improvements. A user study compared and evaluated the effectiveness of these techniques. Preliminary results show that the technique "hover speed," which consists of changing speed to show a tooltip based on the gaze-driven area of interest, was the favourite among the participants. Opinions on the "reference axis" technique were divided. Overall, the interest model improved their performance in both types of tasks, indicating that gaze-driven techniques can support data analytics tasks positively.