Optimal Transport in Python: A Practical Introduction with POT
summary
Claims practical Python code that solves OT problems with concrete scaling: exact solvers run about n^3 log n so handle a few thousand to tens of thousands of points, POT implements ~80 literature variants and cites use in over a thousand papers, regularized Sinkhorn regularization provides smooth, differentiable GPU-backed solvers for gradients, a recent batch GPU solver ships to solve many small problems in parallel, examples include barycenters, graph/brain correspondences and domain adaptation and the library reports working, not just theory.
AI-written summary · about
chapters (10)
- 0:00Optimal transport intro
- 1:04What is optimal transport
- 3:34Optimal transport in ML
- 8:26OT in data science
- 14:24Solving OT with POT
- 17:14Regularized OT solutions
- 18:42Deep learning with OT
- 21:33Acknowledgments
- 22:15POT toolbox overview
- 23:47Q&A on OT
related talks
01h 22mComputational Optimal Transport
15mGraphical Optimal Transport and Its Applications
28mImproving Efficiency in Transportation Networks using Python
39mPractical Optimization for Stats Nerds
09mA Fast Algorithm for Optimal Transport
01h 28mFollow the gradient: an introduction to mathematical optimisation
31mPython Doesn’t Have to Be Slow: Speeding Up a Large-Scale Optimization Algorithm
44mScientific computing with Python: Tools for the solution of continuous problems