Autoencoder for dimensionality reduction — pros and cons.

medium

Answer

$Pros: non-linear, scalable (SGD), transformable (new data \to embedding), can be regularized (denoising, sparse, contractive).$
$Cons: no orthogonality (hard to interpret), no explained-variance ratios, sensitive to hyperparameters, harder than PCA.$
$Rule: for tabular data < 50 dim, PCA is enough.$
$For images / audio / very high dim, autoencoders (especially convolutional) dominate.$

Check yourself — multiple choice

Same as PCA
Non-linear + scalable + transforms new data + regularizable, but no orthogonality/interpretability; PCA enough for small tabular, AE for high-dim / images
Random
Only supervised

Autoencoder DR: non-linear, scalable, less interpretable than PCA.

#dimensionality-reduction#representation-learning

214 interview questions in this topic.