Justin Burton
Emory University, GA·Physics
Can a computer find the physical laws behind a messy, real-world system just by watching how its parts move? Machine learning has rebuilt known equations from simulated data, but it has found few new laws from real experimental data.
- Fields
Physics
Computer Science
- Topics
dusty plasma
complex plasma
plasma physics
- Grants
- 1 known grant, through Jul 2027
What they work on
- Track the 3D motion of charged micron-sized dust particles in a weakly ionized plasma for tens of minutes, using custom scanning laser-sheet tomography.
- Separate external forces from particle-to-particle forces in dusty plasmas dominated by two-body interactions, using machine learning models with the system's symmetries built in.
- Classify the states of dusty plasmas, such as liquid and solid, and find simpler lower-dimensional descriptions of their crowded, fast-changing motion with statistical methods.
- Discover equations of motion directly from experimental particle data, using these simpler descriptions to seed machine learning, where standard approximations break down.
More about the research
Dusty plasmas are a good test. A plasma is a gas in which electrons have been stripped from their atoms. Dust grains in it pick up electric charge and push and pull on each other. Dusty plasmas occur from the Moon's surface to Saturn's rings, and in the lab each grain can be tracked with cameras and lasers.
Why it matters
The work could show how machine learning can pull reliable physical laws out of noisy experimental data, not just simulations. Such methods may later help researchers studying other systems of many interacting parts, such as microorganism swarms, bird flocks or dust in space. It would also give a clearer picture of the forces that control dusty plasmas.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Chuji Wang
Mississippi State University, MS·Physics
Design an optical trapping system that uses focused laser light to pin down single dust grains suspended in a low-temperature plasma and transport them in a controlled way to any position within the plasma device.
Bhuvnesh Bharti
Louisiana State University, LA·Engineering·Early-career
Use large-scale simulations of coupled magnetic, hydrodynamic and contact interactions, plus analytical theory, to derive design rules for how micron-scale magnetic colloids aggregate and move cooperatively in porous media and open three-dimensional fluids.
Qi Yu
University of California-San Diego, CA·Computer Science·Early-career
Develop deep learning models that forecast spatiotemporal dynamics while obeying physical laws, with the goal of simulating ocean currents, traffic and epidemic spread faster than numerical solvers.
The school
Emory University: 70 labs on PI Pages, $1.2B a year on research, 286 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.