Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Autoencoder node saliency: Selecting relevant latent representations

Journal Article · · Pattern Recognition

The autoencoder is an artificial neural network that performs nonlinear dimension reduction and learns hidden representations of unlabeled data. With a linear transfer function it is similar to the principal component analysis (PCA). While both methods use weight vectors for linear transformations, the autoencoder does not come with any indication similar to the eigenvalues in PCA that are paired with eigenvectors. In this work, we propose a novel autoencoder node saliency method that examines whether the features constructed by autoencoders exhibit properties related to known class labels. The supervised node saliency ranks the nodes based on their capability of performing a learning task. It is coupled with the normalized entropy difference (NED). We establish a property for NED values to verify classifying behaviors among the top ranked nodes. Lastly, by applying our methods to real datasets, we demonstrate their ability to provide indications on the performing nodes and explain the learned tasks in autoencoders.

Research Organization:
Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)
Sponsoring Organization:
USDOE National Nuclear Security Administration (NNSA)
Grant/Contract Number:
AC52-07NA27344
OSTI ID:
1491664
Alternate ID(s):
OSTI ID: 1636754
Report Number(s):
LLNL-JRNL--741590; 896098
Journal Information:
Pattern Recognition, Journal Name: Pattern Recognition Journal Issue: C Vol. 88; ISSN 0031-3203
Publisher:
ElsevierCopyright Statement
Country of Publication:
United States
Language:
English

References (17)

Self-monitoring for maintenance of vehicle fleets journal August 2017
Extremely randomized trees journal March 2006
Connectionist learning procedures journal September 1989
Identification of bursts in spike trains journal January 1992
Neural networks and principal component analysis: Learning from examples without local minima journal January 1989
A survey on feature selection methods journal January 2014
Multi-Site Diagnostic Classification of Schizophrenia Using Discriminant Deep Learning with Functional Connectivity MRI journal April 2018
Consensus self-organized models for fault detection (COSMO) journal August 2011
Unsupervised feature extraction with autoencoder trees journal October 2017
A trainable feature extractor for handwritten digit recognition journal June 2007
A novel hybrid CNN–SVM classifier for recognizing handwritten digits journal April 2012
An efficient semi-supervised representatives feature selection algorithm based on information theory journal January 2017
A Survey on Feature Selection journal January 2016
The genomic and transcriptomic architecture of 2,000 breast tumours reveals novel subgroups journal April 2012
Gradient-based learning applied to document recognition journal January 1998
Reducing the Dimensionality of Data with Neural Networks journal July 2006
A Connection Between Score Matching and Denoising Autoencoders journal July 2011

Cited By (3)

Assessment of Autoencoder Architectures for Data Representation book October 2019
Discriminative stacked autoencoder for feature representation and classification journal January 2020
A Robust Deep Learning Approach for Spatiotemporal Estimation of Satellite AOD and PM2.5 journal January 2020

Similar Records

Autoencoder Node Saliency
Software · Mon Nov 20 19:00:00 EST 2017 · OSTI ID:code-16356

Latent Representation Learning for Structural Characterization of Catalysts
Journal Article · Mon Feb 22 23:00:00 EST 2021 · Journal of Physical Chemistry Letters · OSTI ID:1781506