Advancing Interdisciplinary Research: Progress from the Nexus of Data & AI Seed Fund

In 2024, Waterloo.AI launched the Nexus of Data & AI Seed Fund to support emerging interdisciplinary research collaborations across the University of Waterloo. The program, which was developed following the Nexus event held in May, provides $20,000 in seed funding to project teams working at the intersection of data, artificial intelligence, and the university’s Global Futures priorities.

The goal is to enable early-stage research, build new interdisciplinary collaborations, and position teams for major external funding. Several months into the initiative, all seven funded projects have reported strong developments—ranging from new AI models and open-source tools to national grants and foundational conceptual work.

1. Synthetic Data & Federated Learning for Drug Discovery

PIs: Helen Chen, Yaoliang Yu & Anita Layton

Addressing critical challenges in pharmacokinetic research, this team developed two generative AI models—Imagand and xImagand-DKI—to overcome issues with data sparsity in drug discovery datasets. These models generate high-fidelity synthetic data and incorporate molecular and genomic insights, opening pathways for improving downstream biomedical research.

The team has shared their work through presentations at leading venues and continues to expand collaborations with Princess Margaret Cancer Centre, McGill University, and Leiden University. They have submitted a proposal to the DHDP Innovation Fund and continue to explore additional funding opportunities to support future phases of this work.

2. Ethical & Diverse Facial Recognition Technologies

PIs: Lai-Tze Fan & Elliot Creager

This project has generated substantial academic activity, producing multiple publications, presentations, and invited talks on ethical data practices and responsible AI design. The team has also secured a $382,300 SSHRC Insight Grant and a $7,000 Humanities and Social Sciences Endowment Fund, which will support the expansion of their research on data equity, interdisciplinary methodologies, and policy-relevant approaches to facial recognition technologies.

3. Explainable AI for Personalized Cancer Immunotherapy

PIs: Mohammad Kohandel & Amir-Hossein Karimi

This research applies explainable AI techniques to improve predictions of patient response to immunotherapy. The team has developed a preliminary causal graph, integrated algorithmic recourse methods, and adapted probabilistic techniques for use with real and synthetic data. These findings provide a foundation for future models aimed at supporting personalized treatment planning and advancing understanding of biological response patterns in cancer therapies.

4. Wildfire Spread Prediction Using AI

PIs: Mark Crowley & David Del Ray Fernández

This project released a public U-Net wildfire prediction model, with ongoing work to extend these methods to generative modeling of forest ecosystems. Leveraging multi-resolution lidar datasets, the team is developing advanced representations of forest structure to improve the accuracy and generalizability of wildfire spread modeling.

A journal submission is currently in preparation, and continued collaboration with partners in environmental data science is supporting the next stages of research.

5. AI-Assisted Radiation Treatment Quality Assessment

PIs: Houra Mahmoudzadeh & Kate Larson

The team has developed a preliminary model using local radiation oncology data and has already shared early findings at a national conference. Continued work is focused on scaling the model to broader datasets and exploring additional techniques for evaluating and improving radiation therapy planning. Plans for future research development include pursuing new funding opportunities to support model refinement and clinical integration.

6. AI-Enhanced Pilot Training via Eye-Tracking Analytics

PIs: Jian Zhao & Ewa Niechwiej-Szwedo

Working alongside aviation research partners, this team developed and released an open-source eye-tracking analytics toolkit that synchronizes gaze behavior with aircraft instrumentation data. This tool provides new insights into training, expertise development, and human–computer interaction within aviation environments.

The team has also submitted a proposal to the NFRF Transformation 2026 program and is preparing additional funding applications to continue expanding the toolkit and supporting future experimental studies.

7. Lies, Fictions, and LLMs

PIs: Dan Brown & Jennifer Saul

This project investigates how large language models understand and generate lies, fictions, and assertions—questions central to both AI ethics and the philosophy of language. Early work includes conference presentations by team members exploring how AI-generated content interacts with human expectations about truth, storytelling, and communication.

With foundational conceptual work now established, the team plans to pursue SSHRC funding to support the next phase of research.

Accelerating Research

1. Strong Research Progress

Across seven diverse projects, teams have delivered new AI models, open-source tools, conceptual frameworks, and early publications.

2. Publicly Announced Funding Achievements

Notably, the Ethical & Diverse Facial Recognition project secured a $382,300 SSHRC Insight Grant and a $7,000 Humanities and Social Sciences Endowment Fund, underscoring the impact of early-stage interdisciplinary support.

3. Emerging Collaborations

New collaborations have formed across engineering, health, social sciences, philosophy, and aviation—along with partnerships involving Princess Margaret Cancer Centre, McGill University, Leiden University, and industry stakeholders.

4. Foundation for Future Growth

Each project has built momentum that will support future funding applications, publications, and expanded interdisciplinary work.

Looking Ahead

As the Nexus of Data & AI Seed Fund continues to foster new research across Waterloo, upcoming milestones include:

• New publications and prototype releases

• Continued expansion of interdisciplinary collaborations

• Additional external funding proposals

• Deeper integration of research outputs into community, industry, and academic contexts

The progress achieved during this initial phase highlights the value of early-stage funding in catalyzing innovation. Waterloo.AI looks forward to supporting these teams as they continue transforming foundational ideas into impactful research.