Concrete Autoencoders for Differentiable Feature Selection and Reconstruction

TL;DR

The paper introduces the concrete autoencoder, a differentiable, unsupervised method for selecting informative features and reconstructing data, demonstrating superior performance and cost reduction in gene expression analysis.

Contribution

It presents a novel end-to-end differentiable feature selection method using a concrete selector layer integrated into autoencoders, enabling efficient and effective feature subset identification.

Findings

Outperforms state-of-the-art feature selection methods.

Reduces measurement costs by 20% in gene expression datasets.

Successfully reconstructs input data from selected features.

Abstract

We introduce the concrete autoencoder, an end-to-end differentiable method for global feature selection, which efficiently identifies a subset of the most informative features and simultaneously learns a neural network to reconstruct the input data from the selected features. Our method is unsupervised, and is based on using a concrete selector layer as the encoder and using a standard neural network as the decoder. During the training phase, the temperature of the concrete selector layer is gradually decreased, which encourages a user-specified number of discrete features to be learned. During test time, the selected features can be used with the decoder network to reconstruct the remaining input features. We evaluate concrete autoencoders on a variety of datasets, where they significantly outperform state-of-the-art methods for feature selection and data reconstruction. In particular, on a large-scale gene expression dataset, the concrete autoencoder selects a small subset of genes whose expression levels can be use to impute the expression levels of the remaining genes. In doing so, it improves on the current widely-used expert-curated L1000 landmark genes, potentially reducing measurement costs by 20%. The concrete autoencoder can be implemented by adding just a few lines of code to a standard autoencoder.