Yongxin Chen
Georgia Institute of Technology, GA·Engineering
How do you steer a whole crowd rather than a single object? Imagine a swarm of robots, a population of biological agents, or the particles in a manufacturing process.
- Fields
Engineering
Mathematics & Statistics
Computer Science
- Topics
distribution control
stochastic optimal control
optimal mass transport
- Grants
- 1 known grant, through Jul 2028
What they work on
- Develop theory and algorithms for controlling probability distributions of agents whose dynamics are nonlinear in state and non-affine in control, with optimality guarantees under hard deadlines.
- Design distribution-control methods for populations whose agents interact nonlocally, such as robotic or biological swarms, using tools from optimal mass transport and Schrödinger bridges.
- Build robust, scalable, gridless algorithms that control distributions beyond covariance, with or without process noise and under dynamic constraints, with guarantees on performance.
More about the research
The goal is to shape the overall spread of states across the population, which mathematicians call a probability distribution. Existing methods fall short in practice. They ignore realistic nonlinear behavior of individual agents, interactions between agents, statistical errors and constraints. The same problem appears in machine learning, where generative models must precisely control distributions of data.
Why it matters
Reliable methods for steering whole distributions could support precision manufacturing of materials with desired properties. They could also help coordinate large robotic or biological swarms. Because the same mathematics underlies generative diffusion models, the work could also improve how machine learning systems control data distributions.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Huibin Zhou
Yale University, CT·Mathematics & Statistics
Determine whether the iterative rectified flow algorithm, which turns high-dimensional density estimation into repeated regression, achieves optimal density and transport estimation in just one step of regression.
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.
Yusu Wang
University of California-San Diego, CA·Computer Science
Explain why trained flow-based and diffusion generative models generalize better than the theoretically optimal solution predicts, using geometry, ODE analysis, manifold learning and deep learning theory, to enable controlled generation of novel samples.
The school
Georgia Institute of Technology: 50 labs on PI Pages, $1.5B a year on research, 572 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.