The Deep Kernelized Autoencoder
This work addresses the challenge of controlling similarity in embeddings for machine learning practitioners, offering an incremental improvement over existing autoencoder methods.
The paper tackles the problem of learning effective data representations in autoencoders by aligning inner products between codes with a kernel matrix, resulting in similarity-preserving embeddings with competitive performance in classification and visualization tasks, and emulating kernel PCA at lower computational cost.
Autoencoders learn data representations (codes) in such a way that the input is reproduced at the output of the network. However, it is not always clear what kind of properties of the input data need to be captured by the codes. Kernel machines have experienced great success by operating via inner-products in a theoretically well-defined reproducing kernel Hilbert space, hence capturing topological properties of input data. In this paper, we enhance the autoencoder's ability to learn effective data representations by aligning inner products between codes with respect to a kernel matrix. By doing so, the proposed kernelized autoencoder allows learning similarity-preserving embeddings of input data, where the notion of similarity is explicitly controlled by the user and encoded in a positive semi-definite kernel matrix. Experiments are performed for evaluating both reconstruction and kernel alignment performance in classification tasks and visualization of high-dimensional data. Additionally, we show that our method is capable to emulate kernel principal component analysis on a denoising task, obtaining competitive results at a much lower computational cost.