Supervised Random Projections with Light

Python implementation of supervised PCA, supervised random projections, and their kernel counterparts.

Supervised Random Pojections (SRP) is the work of Amir-Hossein Karimi, Alexander Wong, and Ali Ghodsi. It is a fast approximation of the Supervised PCA algortithm for dimensionality reduction. It also has a nonlinear version, Kernel SRP (KSRP).

This repository provides a unified implementation of SPCA, KSPCA, SRP and KSRP. They are implemented as scikit-learn transformers, and can therefore be used exactly like scikit-learn's PCA and KPCA. Moreover, SRP and KSRP can be performed using a LigthOn Optical Processing Unit (OPU).

• dimreduc.py contains the implementations of the algorithms;
• load_data.py contains utilities to load the datasets used in the original paper (XOR, Spirals, Sonar and Ionosphere);
• sonar_viz.py shows how to use this code for visualizing the Sonar dataset.

The Ionosphere and Sonar dataset come from the UCI repository. They are tiny, so I included them in the data folder for convenience.

Access to Optical Processing Units

To request access to LightOn Cloud and try our photonic co-processor, please visit: https://cloud.lighton.ai/

For researchers, we also have a LightOn Cloud for Research program, please visit https://cloud.lighton.ai/lighton-research/ for more information.