Taniya Kapoor
Wageningen University & Research, Delft University of Technology, University of Lille, South Asian University, ETH Zurich, University of Delhi
0000-0002-6361-446X
About
I am an Assistant Professor in the Artificial Intelligence Group at WUR. My research interests lie broadly in Machine Learning for the Sciences and Science for Machine Learning. In particular, I focus on developing ML models that integrate physical knowledge such as partial differential equations, ordinary differential equations, and physical laws, to tackle global challenges.
I obtained my PhD at TU Delft, where I worked on physics-informed machine learning for structural engineering applications under the supervision of Dr. Hongrui Wang, Dr. Alfredo Núñez, and Prof. Rolf Dollevoet. Following the completion of my PhD, I was awarded the AI Schmidt Postdoctoral Fellowship at the University of Oxford, to develop engineering-informed foundation models for sustainable structural design. Prior to that, I worked at ETH Zürich (CAMLab), on physics-informed neural networks for fully nonlinear PDEs.
Employment
-
Wageningen University & Research Assistant Professor2025 - Present
-
Delft University of Technology Postdoc Researcher2024 - 2025
-
Delft University of Technology PhD Employee2021 - 2024
-
ETH Zurich Master Intern2021 - 2021
-
Wageningen University & Research
Education
-
University of Lille Master of Science in High Performance Scientific Computing2020 - 2021
-
South Asian University Master of Science in Applied Mathematics2017 - 2019
-
University of Delhi Bachelors of Science in Mathematics (Hons.)2014 - 2017
Projects & Funding
Projects & funding information is unavailable.
Publications (20)
- A unified causality-enhanced separable physics-informed neural network for predicting beam and plate dynamics Save
- Curvature-aware dynamic precision approach for physics-informed neural networks Save
- Neuro-symbolic operator for interpretable and generalizable characterization of complex piezoelectric systems Save
- Curvature-aware dynamic precision approach for physics-informed neural networks Save
- Generalizable models of magnetic hysteresis via physics-aware recurrent neural networks Save
- Neural Differential Equation-Based Two-Stage Approach for Generalization of Beam Dynamics Save
- Extrapolating Plate Dynamics Through Neural Differential Equations Save
- Separable Physics Informed Neural Networks Save
- Physics-informed machine learning for moving load problems Save
- Transfer learning for improved generalizability in causal physics-informed neural networks for beam simulations Save