Bridging the Gap: From Computational Physics, to Physics-informed Machine Learning, to Data-driven Scientific Discovery

This is the fourth of several introductory mini-symposia from the DSI Emergent Data Sciences program for the Emergent Data Science program, Bridging the Gap: From Computational Physics, to Physics-informed Machine Learning, to Data-driven Scientific Discovery. The program aims to bring together experts in numerical simulation and data science to explore the intersection between, and bridge the gap separating, physics-based models grounded in first principles and data-driven models based on machine learning techniques.

October 28, 2026
1:10-1:15
Opening Remarks
1:15-1:45
Talk title: TBC
David Zingg, Institute for Aerospace Studies, Faculty of Applied Science and Engineering, University of Toronto
1:45-2:15
Deep Learning with Learnable Product-Structured Activations
Saanjali Maharaj, Institute for Aerospace Studies, Faculty of Applied Science and Engineering, University of Toronto
2:15-2:45
Talk title: TBC
Speakers: TBC
2:45
Closing remarks

Speakers

Saanjali Maharaj
Institute for Aerospace Studies, Faculty of Applied Science and Engineering, University of Toronto

Maharaj is a 4th year PhD candidate in the Decision Analytics for Computational Engineering (DACE) Research Group, led by Professor Prasanth B. Nair. Her work focuses on the development of novel neural network architectures designed to overcome the mathematical and structural limitations of standard machine learning models. Her current research explores the integration of low-rank tensor decomposition theory into deep learning to create highly expressive, parameter-efficient architectures. Saanjali is particularly interested in designing adaptive models that mitigate the curse of dimensionality and spectral bias in complex systems. She has applied these architectures to a diverse range of computationally intensive challenges, including scientific machine learning for partial differential equations (PDEs), inverse problems such as sparse-view medical image reconstruction, and high-fidelity continuous image and audio signal representation.

David Zingg
Institute for Aerospace Studies, Faculty of Applied Science and Engineering, University of Toronto

Prof. Zingg research areas include aerodynamics, computational fluid dynamics, in particular high-order methods with the summation-by-parts property, and aerodynamic shape optimization. His current research is concentrated on applying aerodynamic shape optimization to the design of unconventional low-drag aircraft configurations motivated by the need to reduce greenhouse gas emissions from aircraft.

October 28, 2026

Data Sciences Institute,
Seminar room
10th floor,
700 University Avenue