SUDS project

Oct 02 2026

Developing Sleep Behavioural Model using Transfer Learning Approaches

Research description:

Improving in recent years is the exploration of mobile health (mHealth) to empower users with better sleep hygiene. The adverse sequelae of poor sleep hygiene can affect a person’s physical fitness and emotional and mental wellness; thus, its significance has spurred much technological advancement, ranging across a broad spectrum of smart and wearable devices to support personalized sleep health. Yet, the continued realization of mHealth sleep solutions that strategically integrate with society and the healthcare environment has seen more fragmentation than coordination. This project tackles a growing concern in developing ML-driven sleep behavioral models, often including bias learning from the studied population. Indeed, prior work has reported differences in sleep model performance between children and older adults. Transfer learning, in general, has shown feasibility in achieving subject-independent classification using pre-training and fine-tuning paradigms in many activity recognition applications. In the same manner, students will investigate the efficacy of such techniques for sensor-driven mobile data predicting sleep measures.

Through this project, the SUDS Scholar will:

Learn to identify and develop potential ML applications for sleep detection using publicly available and user-study datasets
Learn to design and implement human-interpretable results and output created by the ML models.

Year: 2024

Researcher:
Camellia Zakaria, Dalla Lana School of Public Health, Faculty of Applied Science and Engineering, University of Toronto

Student:
Tae-Kyeong Kim, University of Toronto

Through SUDS, undergraduate students engage in hands-on research focused on data sciences and AI methodology applications.

Learn more

Oct 02 2026

Development of deep learning model for accelerating chemical reaction discovery

Research description:

Fast and accurate computational modeling of chemical properties and structures has immense potential to accelerate discovery in various fields, including drug design and catalysis. In this context, predicting transition state (TS) structures of chemical reactions that cannot be obtained experimentally offers a powerful way to gather mechanistic details and generate energy profiles. Unfortunately, most traditional computational chemistry methods for predicting TS structures are still very costly to expedite high-throughput applications. Deep learning (DL) can potentially substitute traditionally used expensive quantum mechanical (QM) methodologies in chemistry to predict these structures by offering to generate fast and accurate mathematical models better suited to everyday computers. In this project, the student will further explore this promising avenue and utilize a chemical reaction data set being currently generated in-house. They will utilize it to develop a novel graph neural network (GNN) based model for predicting the highly desired transition state structures at an unprecedented speed and demonstrating the acceleration provided via DL. The research aims at providing a novel way to cut down the computational cost and manual intervention associated with TS structure prediction. Such developments hold immense potential to advance computational chemistry and accelerate high-throughput applications like drug discovery and catalysis. The student will assist in solution design, development of research code (with PyTorch, Keras, Tensorflow, and Python), deployment and running on HPCs. They will help with the generation of reference chemical property datasets. They will then be engaged mainly in designing graph convolutional neural network architecture for structure predictions. The proposed research will be coordinated by the Supervisor and a postdoctoral fellow with experience in running interdisciplinary collaborations. A key responsibility will be to provide one-on-one support to the student, which includes guidance on interdisciplinary method development, data analysis, result interpretation, and effective research communication in the form of published articles and presentations.

Year: 2024

Researcher:
Hans-Arno Jacobsen, Edward S. Rogers Sr. Department of Electrical and Computer Engineering, Faculty of Applied Science & Engineering, University of Toronto

Student:
Shivesh Prakash, University of Toronto

Through SUDS, undergraduate students engage in hands-on research focused on data sciences and AI methodology applications.

Learn more

Oct 02 2026

Development of deep learning solutions for 4D image segmentation of spatial omics data

Research description:

How one genome generates a large diversity of cell types, each with unique spatiotemporal gene expression patterns and physiological roles, is an enduring fundamental question in cell and developmental biology. To understand the genome structure-function relationships, it is not sufficient to know the genome sequence and local epigenetic features – we must also consider the large-scale physical architecture of entire chromosomes and their positioning within the nucleus in space and time (4D). Our broad objective is to understand how the entire genome is organized in complex multicellular systems, and how this organization influences the genome’s functional output (Sawh et al., Mol Cell 2020; Sawh and Mango. Current Opinion in Genetics & Development 2022). In the current opportunity, a SUDS scholar will extend our methods of traditional watershed 3D image segmentation to extract quantitative chromosome conformation information from C. elegans embryo spatial omics data. With a large amount of 3D segmented ground truth data in hand, the applicant will develop a threshold-free deep neural network approach to accurately segment anisotropic cell, nuclear, and chromosome objects in C. elegans embryos over developmental time. The position can be in-person, remote, or hybrid depending on the preference of the candidate.

The student will work closely with graduate students and myself, to use and refine 3D semantic image segmentation algorithms on fluorescence images. The student will refine a pre-trained neural network model (e.g. Cellpose 2.0) using already available ground-truth data to develop custom models for C. elegans cell, nuclei, and chromosome volumes.

Year: 2024

Researcher:
Ahilya Sawh, Department of Biochemistry, University of Toronto

Students:
Faisal Shaik, University of Toronto
Waleed Adel Alsarhani, King Abdullah University of Science & Technology

Through SUDS, undergraduate students engage in hands-on research focused on data sciences and AI methodology applications.

Learn more

Oct 02 2026

Diffusion Policies for Discrete and Continuous Actions

Research description:

Advancements in reinforcement learning have paved the way for more sophisticated policies for decision-making in complex environments. This project delves into the exploration of diffusion policies, a class of algorithms that leverage stochastic processes to model the decision-making process in both discrete and continuous action spaces. The primary objective is to design, implement, and evaluate diffusion policies for a range of applications, showcasing their adaptability and effectiveness in diverse scenarios. The student will be responsible for implementing the method, conducting experiments, and compiling the results into a paper.

Year: 2024

Researcher:
Florian Shkurti, Department of Mathematical and Computational Sciences, University of Toronto Mississauga

Student:
Jierui Zhu, University of Toronto

Through SUDS, undergraduate students engage in hands-on research focused on data sciences and AI methodology applications.

Learn more

Oct 02 2026

Data science and global health: investigating the relationship between disease and data science methodology

Research description:

Novel statistical and data science methodology developments are often motivated by the increasingly complex data collected for research studies that involve a disease or health outcome. For example, methodological studies motivated by breast cancer or Alzheimer’s disease are common. However, whether the effort for methodological development is appropriately being used for diseases that affect the global population the most is unknown.

In 2020, Lancet published the updated global burden of 369 diseases and injuries in 204 countries and territories. For each top 25 disease, using an automated systematic literature review, we will identify all published methodological research motivated by the disease.

Then we will:

Assess the relationship between the common global diseases and the common diseases that motivate methodological research.
Identify global diseases that are being “neglected” by methodologists.

The student will learn how to conduct an automated systematic literature review, text analysis, produce professional tables, figures, and graphics. The student will have the opportunity to be a co-author of a paper in an academic journal. The R statistical programming language will be primarily used, but other programming languages may be considered based on the student’s proficiency.

Year: 2024

Researcher:
Aya Mitani, Dalla Lana School of Public Health, University of Toronto

Student:
Bilin Nong, University of Toronto

Through SUDS, undergraduate students engage in hands-on research focused on data sciences and AI methodology applications.

Learn more