PCA and autoencoders Principal component analysis (PCA) is an example of dimensionality reduction. Autoencoders generalize the idea to non-linear transformations. The interactive Jupyter notebook let's you train such autoencoders yourself, and see the results.