General
Scalable Learning from Probability Measures with Mean Measure Quantization
arXiv stat.ML0 views
arXiv:2502.04907v5 Announce Type: replace Abstract: We consider statistical learning problems in which data are observed as a set of probability measures. Optimal transport (OT) is a popular tool to compare and manipulate such objects, but its computational cost becomes prohibitive when the measures have large support. We study a quantization-based approach in which all input measures are approximated by $K$-point discrete measures sharing a common support. We establish consistency of the resulting quantized measures. We further derive convergence guarantees for several OT-based downstream tasks computed from the quantized measures. Numerical experiments on synthetic and real datasets demonstrate that the proposed approach achieves performance comparable to individual quantization while substantially reducing runtime.
Source:https://arxiv.org/abs/2502.04907PreviousHigh-Dimensional Tail Index RegressionNextA Cherry-Picking Approach to Large Load Shaping for More Effective Carbon Reduction
Related posts
Show HN: Type Fighter 2
0Hacker News
Evidence (YC S21) Is Hiring
0Hacker News
Helvetica: The Typeface Everyone Respects but Nobody Questions
0HackerNoon
How Claude Code Agents Accidentally Delete Files—and How to Stop Them
0HackerNoon
Claude-Mem Tested: Perfect Recall, but at a Cost of 5,200 Tokens Per Request
0HackerNoon
A Beginner's Guide to Classes and Objects in C++
0HackerNoon