gplvm-maps-latent-to-data-opposite-kpca
IN premise — summaries/2026/08/24/wiki-Nonlinear_dimensionality_reduction-chunk-1.md
Created 2026-08-24T17:11:20+00:00
GPLVM (Gaussian Process Latent Variable Model) maps from latent space to data space, the opposite direction from Kernel PCA which maps from data space to latent space.
Summary
These two methods solve mirror-image problems: one takes a compact internal representation and produces the data you observe, while the other takes the observed data and compresses it down into a compact representation. Getting the direction wrong matters because it changes what you can actually use each tool for — one is a generative model, the other is a compression tool, and swapping them would lead to incoherent pipelines.