Tamer Zaki
Johns Hopkins University, MD·Engineering
Many problems in engineering and physics are hard to simulate not because the physics is mysterious but because the equations involve too many variables at once. Picture a liquid jet breaking into an enormous number of droplets, each carried and deformed by a swirling background flow; the same mathematics describes airborne particles carrying pathogens through a room.
- Fields
Engineering
Computer Science
Mathematics & Statistics
- Topics
evolutional deep neural network
partial differential equations
high-dimensional problems
- Grants
- 1 known grant, through Dec 2026
What they work on
- Develop evolutional deep neural network (EDNN) algorithms that solve high-dimensional partial differential equations by evolving network parameters according to the governing physics, without relying on training data.
- Determine optimal network architectures for EDNN representations of high-dimensional solutions, and assess accuracy against benchmark high-fidelity datasets.
- Build dynamic adaptivity into the EDNN solution so the representation adjusts as the solution evolves localized or sharp features.
- Establish scalability of the EDNN algorithms for massively parallel computation on large high-dimensional problems.
More about the research
As the number of particles grows, the cost of solving the governing equations explodes. Faster methods for these high-dimensional partial differential equations would help across fluid mechanics, physics, medicine and public health.
Why it matters
Simulations of high-dimensional problems—droplet-laden flows, particle-laden air, and other many-variable systems—are currently limited by computational cost. Training-data-free neural network solvers could make such predictions affordable, which matters for engineering design and for questions such as airborne transmission of pathogens that bear on medicine and public health.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Burak Aksoylu
Texas A&M University-San Antonio, TX·Mathematics & Statistics
Use machine learning to improve iterative local-to-nonlocal coupling, so that peridynamics models only the region where fracture occurs and cheaper local models handle the rest.
Yiran Wang
University of Alabama, AL·Mathematics & Statistics
Develop hybrid reduced order models in which neural networks add basis functions to precomputed multiscale ones, guided by stable Petrov-Galerkin formulations and residual-based error indicators.
Haomin Zhou
Georgia Institute of Technology, GA·Mathematics & Statistics
Establish parameterized Wasserstein sub-manifolds, which represent evolving probability distributions through a neural network's parameters via a push-forward map and pullback metric, and develop their mathematical foundations.
The school
Johns Hopkins University: 100 labs on PI Pages, $4.1B a year on research, 622 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.