Language selection

Government of Canada / Gouvernement du Canada

Search

PCA-enhanced autoencoders for nonlinear dimensionality reduction in low data regimes

From National Research Council Canada

Download
  1. (PDF, 2.2 MiB)
DOIResolve DOI: https://doi.org/10.21428/594757db.05a13011
AuthorSearch for: 1; Search for: 1; Search for: ; Search for: 2; Search for: 2; Search for: 2; Search for: 2
Affiliation
  1. National Research Council Canada. Digital Technologies
  2. National Research Council Canada. Advanced Electronics and Photonics
FormatText, Article
ConferenceThe 36th Canadian Conference on Artificial Intelligence (Canadian AI 2023), June 5-9, 2023, Montréal, Québec
Physical description12 p.
Subjectdimensionality reduction; autoencoders; principal component analysis (PCA); limited datasets
Abstract
Many scientific domains, such as nanophotonic design, gene expression, and materials design, are limited by high costs of acquiring data. This data is often intrinsically low-dimensional, nonlinear, and benefits from dimensionality reduction. Autoencoders (AE) provide nonlinear dimensionality reduction but are typically ineffective for low data regimes. Principal Component Analysis (PCA) is data-efficient but limited to linear dimensionality reduction. We propose a technique that harnesses the benefits of both methods by using PCA to initialize an AE. The proposed approach outperforms both PCA and standard AEs in low-data regimes and is comparable to the best of either of the two in other scenarios.
Publication date
PublisherCanadian Artificial Intelligence Association
Licence
In
  • Proceedings of the 36th Canadian Conference on Artificial Intelligence (5 June 2023).
LanguageEnglish
Peer reviewedYes
Export citationExport as RIS
Report a correctionReport a correction (opens in a new tab)
Record identifierddd45128-1b17-43b4-9c55-7fb4dc3c954f
Record created2023-08-23
Record modified2023-08-28
Date modified: