Learning Paths for Dynamic Measure Transport: A Control Perspective
By: Aimee Maurais, Bamdad Hosseini, Youssef Marzouk
Potential Business Impact:
Finds better ways to move data for computers.
We bring a control perspective to the problem of identifying paths of measures for sampling via dynamic measure transport (DMT). We highlight the fact that commonly used paths may be poor choices for DMT and connect existing methods for learning alternate paths to mean-field games. Based on these connections we pose a flexible family of optimization problems for identifying tilted paths of measures for DMT and advocate for the use of objective terms which encourage smoothness of the corresponding velocities. We present a numerical algorithm for solving these problems based on recent Gaussian process methods for solution of partial differential equations and demonstrate the ability of our method to recover more efficient and smooth transport models compared to those which use an untilted reference path.
Similar Papers
Machine-Learned Sampling of Conditioned Path Measures
Machine Learning (Stat)
Helps computers learn without needing examples.
Optimal and Diffusion Transports in Machine Learning
Optimization and Control
Makes AI learn and create better using math.
A Geometric Approach to Optimal Experimental Design
Machine Learning (Stat)
Finds the best way to test ideas.