Please join the Computer Science Department for the upcoming thesis defense:

Presenter: Rustem Kakimov

Thesis title: Upward Book Embeddings of DAGs: Constraint-Based Methods and Embeddability Analysis

Abstract: The k-page upward book embedding (kUBE) problem is a fundamental challenge in graph theory with applications in circuit layout, scheduling, and hierarchical visualization. Despite its relevance, the problem—particularly for k ≥ 2—remains underexplored. This thesis develops practical methods for solving kUBE and conducts a detailed investigation of how graph structural properties influence upward embeddability.

We first propose a Boolean satisfiability (SAT) encoding, SAT-1, that extends existing k-page book embedding techniques to the general kUBE setting. For the special case of k = 2 (2UBE), we introduce SAT-2, a more compact SAT encoding exploiting the fixed number of pages, and a constraint programming (CP) model as an alternative formulation. Empirical evaluation shows that SAT solvers consistently outperform CP, with SAT-2 achieving up to 40% faster runtimes on large instances and up to 30× speedups on hard instances from the North dataset compared to SAT-1.

Beyond solving efficiency, we systematically analyze how upward book embeddability depends on structural parameters such as the edge-to-vertex ratio (m/n). Through exhaustive enumeration and sampling, we identify sharp phase transition phenomena across different values of k (up to k = 6) and model the phase transition threshold as a function of graph size and page count using a power-law relationship, providing the first quantitative characterization of this phenomenon.


Committee Members:
Dr. Xing Tan (supervisor, committee chair), Dr. Ruizhong Wei, Dr. Kai Huang (McMaster University)

Please contact grad.compsci@lakeheadu.ca for the Zoom link. Everyone is welcome.

Have you ever wondered, why lawns? Dementia Gardens is proud to have Dr. Harvey Lemelin presenting at our next speaker series event, "Replacing Colawnization Through Community." Dr. Harvey Lemelin is a professor at Lakehead University in the Outdoor Recreation and Leisure department. He will be presenting an overview of the strange leisure legacy of lawns in our society. We hope to see you there!

The Collaborative Doctor of Veterinary Medicine Program (CDVMP) Information Session is a casual, drop-in event where members of the university community can learn more about the program, explore its development, and ask questions. Refreshments will be provided. 

For more information or questions, please contact Dr. Michel S. Beaulieu, Associate Vice-Provost (Academic) at avp.academic@lakeheadu.ca

Please join the Computer Science Department for the upcoming thesis defense:

Presenter: Arvind Chidambaram Boominathan

Thesis title: Integrating Multi-omics Data via Latent Space Construction for Breast and Bladder Cancer Analysis

Abstract: Cancer remains one of the most complex and heterogeneous diseases, driven by intricate interactions across genetic, epigenetic, and transcriptional landscapes. Accurately understanding and predicting tumor characteristics, such as Tumor Mutational Burden (TMB), is critical for effective diagnosis, prognosis, and personalized treatment strategies. This research aims to address inherent challenges in integrating high-dimensional, heterogeneous multi-omics datasets—including DNA methylation, gene expression, and Copy Number Alteration (CNA)—specifically for bladder and breast cancer analysis, by building a shared latent space that captures and preserves meaningful cross-omics representations. Some of these challenges include data imbalance, dimensionality, modalityspecific noise, and complex non-linear biological interactions.

To overcome these obstacles, this thesis proposes constructing a shared latent space through advanced deep-learning approaches by utilizing Deep Multiset Canonical Correlation Analysis (DMCCA) and Graph Attention Networks (GATs). The shared latent space methodology provides a unified representation capturing crucial and intricate biological interactions across various omics modalities, as a result giving improved predictive accuracy for TMB classification. Attention mechanisms further refine this integration by dynamically focusing on the most relevant relational patterns within multiomics data, enhancing the model’s ability to capture biological interactions between genes, pathways, and patient profiles. In addition, this study utilizes oversampling techniques—mainly the Synthetic Minority Oversampling Technique (SMOTE)—to offset data imbalance among TMB classes and menopausal status groups. As compared to baseline supervised machine learning models such as Logistic Regression (LR), Artificial Neural Network (ANN), and Tabular Transformer, the new GAT model with shared latent space training performed better by achieving an AUC of 0.76 and accuracy of 76.1% for BRCA, whereas that of BLCA was 0.73 with an accuracy of 65.3%, thereby establishing the usefulness of multi-omics integration through shared latent space learning.

Committee Members:
Dr. Abedalrhman Alkhateeb (supervisor, committee chair), Dr. Saad Bin Ahmed, Dr. Abdulsalam Yassine (Software Engineering)

Please contact grad.compsci@lakeheadu.ca for the Zoom link. Everyone is welcome.

Lakehead International offers drop-in sessions every Monday from 2 to 4 p.m. for students to ask any immigration-related questions. You can join the session both online and in person. Please note our session is first come first serve basis.

If you choose an in-person session, please visit OA 1031.

If you choose an online session, please click this link: https://lakeheadu.zoom.us/j/91593486944

The Lakehead Town & Gown brings Orillia faculty together with members of the surrounding community. Meetings occur monthly to discuss topics of broad interest—such as AI, education, loneliness, and citizenship—and have easy but engaged conversation. The group is co-run by Ryan McVeigh, Associate Professor of Interdisciplinary Studies, and Anderson Charters, retired publisher and champion of Lakehead’s Orillia campus.

The next session will occur on Friday, May 2, from 12–2:30 p.m. in OA 2005. The topic for discussion is the future of Canada. 

Title: "Does Anthropogenic Disturbance Enhance Breeding Habitat at a Poleward Range Edge?"

Supervisory Committee:
Dr. Adam Algar (Supervisor)
Dr. Cody Dey (adjunct, Lakehead University)
Dr. Julie Yee-Law (University of Ottawa)

Please join Abby in ATAC 3004 or contact Heather at biology@lakeheadu.ca for the Zoom link.

Candidate: Georgina Tough

Title: "Investigating the use of paper mill residuals as agricultural soil amendments in Thunder Bay, Ontario”

Supervisory committee:

Dr. Amanda Diochon (Supervisor)
Dr. Kam Leung
Dr. Tarlock Singh Sahota
Dr. Nathan Basiliko (External)

Please join Georgina in ATAC 3004 or contact Heather at biology@lakeheadu.ca for the Zoom link.

Everyone is invited to attend the following candidate's public presentation for the Department of Biology Plant Biodiversity and Evolution position at Lakehead University.

Dr. Amanda De La Torre
Research Seminar - "How will conifers adapt to climate change? A journey through time and space based on genomic analyses”

Please contact biology@lakeheadu.ca for the Zoom link. All are welcome to attend.

Everyone is invited to attend the following candidate's public presentation for the Department of Biology Plant Biodiversity and Evolution position at Lakehead University.

Dr. Stephanie Rivest
Research Seminar - "Impacts of human activities on insect-plant communities"

Please contact biology@lakeheadu.ca for the Zoom link. All are welcome to attend.