Skip to content

Latent optimal transport (LOT) for low rank transport and clustering

License

Notifications You must be signed in to change notification settings

uldyssian2008/latentOT

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

10 Commits
 
 
 
 
 
 
 
 

Repository files navigation

latentOT

Latent optimal transport (LOT) for simultaneous aligning and clustering

Overview

Latent optimal transport is a low-rank distributional alignment technique. It is suitable for data admitting clustering structure, where alignment is based on a clustering level to make a transport more robust to outliers and noise. Users could customize their cost matrix to fit their clustering strategies. The algorithm requires two number of anchors to be pre-specified. THe numbers naturally correpsond to the numbers of clusters for the source and target.

Dependencies

sklearn, numpy, matplotlib, POT: Python Optimal Transport Rémi Flamary and Nicolas Courty, POT Python Optimal Transport library, Website: https://pythonot.github.io/, 2017

Usage

The code contains a Python implementation on LOT.

lOT.py contains the module for LOT.

lot_mnist_demo.ipynb provides a simple demo on distributional alginments for pre- and post-droupout of MNIST digits.

About

Latent optimal transport (LOT) for low rank transport and clustering

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Jupyter Notebook 98.5%
  • Python 1.5%