Structuring Autoencoders
This work addresses the challenge of discovering semantic structure in raw data for applications such as classification with limited labels, but it is incremental as it builds on existing autoencoder methods.
The paper tackles the problem of learning structured low-dimensional representations from data by proposing Structuring AutoEncoders (SAE), which enhance traditional autoencoders with weak supervision to form a structured latent space, resulting in more efficient data representation for tasks like classification with sparsely labeled data, as demonstrated on benchmark datasets including MNIST, Fashion-MNIST, DeepFashion2, and 3D human shapes.
In this paper we propose Structuring AutoEncoders (SAE). SAEs are neural networks which learn a low dimensional representation of data which are additionally enriched with a desired structure in this low dimensional space. While traditional Autoencoders have proven to structure data naturally they fail to discover semantic structure that is hard to recognize in the raw data. The SAE solves the problem by enhancing a traditional Autoencoder using weak supervision to form a structured latent space. In the experiments we demonstrate, that the structured latent space allows for a much more efficient data representation for further tasks such as classification for sparsely labeled data, an efficient choice of data to label, and morphing between classes. To demonstrate the general applicability of our method, we show experiments on the benchmark image datasets MNIST, Fashion-MNIST, DeepFashion2 and on a dataset of 3D human shapes.