Matching component analysis for transfer learning

DOI10.1137/19M1285123MaRDI QIDQ5027030zbMATH OpenOpenAlexFDO

Authors Charles Clum, Dustin G. Mixon, Theresa Scarnati

Publication date 3 February 2022

Published in SIAM Journal on Mathematics of Data Science (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1909.01797

zbMATH Keywords

classificationdimensionality reductiontransfer learningdomain adaptation

Mathematics Subject Classification ID

Numerical mathematical programming methods (65K05) Learning and adaptive systems in artificial intelligence (68T05) Pattern recognition, speech recognition (68T10) Applications of mathematical programming (90C90) Stochastic programming (90C15)

Abstract: We introduce a new Procrustes-type method called matching component analysis to isolate components in data for transfer learning. Our theoretical results describe the sample complexity of this method, and we demonstrate through numerical experiments that our approach is indeed well suited for transfer learning.

Recommendations

• Subspace projection techniques in feature transfer learning
• On Universal Transfer Learning
• Transfer learning via multi-view principal component analysis
• Adaptive transfer learning
• Analysis on methods to effectively improve transfer learning performance

Cites work

• AlexNet
• An introduction to matrix concentration inequalities
• Deep learning: an introduction for applied mathematicians
• High-dimensional probability. An introduction with applications in data science
• Matrix concentration inequalities via the method of exchangeable pairs
• Sparse representation based Fisher discrimination dictionary learning for image classification
• SqueezeFit: Label-Aware Dimensionality Reduction by Semidefinite Programming
• Visualizing data using t-SNE

Cited in

(6)

• Microtexture region segmentation of eddy current testing data using a structural prior
• Transfer learning via multi-view principal component analysis
• Lie PCA: density estimation for symmetric manifolds
• Balanced joint maximum mean discrepancy for deep transfer learning
• A Matrix Factorization & Clustering Based Approach for Transfer Learning
• Analysis on methods to effectively improve transfer learning performance

Describes a project that uses

Uses Software

• AlexNet
• pix2pix
• CycleGAN
• t-SNE
• ImageNet
• MNIST

This page was built for publication: Matching component analysis for transfer learning

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5027030)