Dimensionality Reduction Paper - Experiment

Description: Below you can find the GitHub repository associated with my 2023 REU Paper. In this repository you will find:

• A NumPy implementation of a simple MLP (the MLP is then trained on MNIST up to 85% accuracy)
• Comparison of gradient-descent PCA, baseline PCA, and random projection
• Comparison of standard Laplacian Eigenmaps with ALE and “VRLE” (as outlined in the paper)

Link to Repo

This accompanying project allowed me to apply the theory developed in my paper and taught me reformulating math into code takes many levels of optimization (and can often always be further optimized, as is certainly the case here).