Concrete Autoencoders for Differentiable Feature Selection and Reconstruction

27 Jan 2019  ·  Abubakar Abid, Muhammad Fatih Balin, James Zou ·

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.

PDF Abstract
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
General Classification Activity CAE Accuracy 42 # 1
General Classification Selection Coil-20 CAE Accuracy 58.6 # 1
General Classification Fashion-MNIST CAE Accuracy 67.7 # 1
General Classification ISOLET CAE Accuracy 68.5 # 1
General Classification Mice Protein CAE Accuracy 13.4 # 1
General Classification MNIST CAE Accuracy 90.6 # 1