Evaluating Autoencoders for Parametric and Invertible Multidimensional Projections
This work addresses a specific need in data visualization for researchers and practitioners by enabling efficient embedding and data generation, though it is incremental as it builds on existing autoencoder and projection methods.
The paper tackled the problem of creating parametric and invertible multidimensional projections simultaneously, which had not been explored before, and found that autoencoders with a customized loss function produce smoother projections than feed-forward neural networks, offering user control over smoothing.
Recently, neural networks have gained attention for creating parametric and invertible multidimensional data projections. Parametric projections allow for embedding previously unseen data without recomputing the projection as a whole, while invertible projections enable the generation of new data points. However, these properties have never been explored simultaneously for arbitrary projection methods. We evaluate three autoencoder (AE) architectures for creating parametric and invertible projections. Based on a given projection, we train AEs to learn a mapping into 2D space and an inverse mapping into the original space. We perform a quantitative and qualitative comparison on four datasets of varying dimensionality and pattern complexity using t-SNE. Our results indicate that AEs with a customized loss function can create smoother parametric and inverse projections than feed-forward neural networks while giving users control over the strength of the smoothing effect.