Minimax and Adaptive Transfer Learning for Sparse Canonical Correlation Analysis
Abstract
We develop a transfer learning framework for high-dimensional sparse canonical correlation analysis (CCA) with multiple heterogeneous source datasets. Our goal is to improve estimation of the target canonical coefficient matrices by borrowing information from related sources. We characterize source-target similarity through discrepancies in the joint covariance structure and establish minimax optimal rates under both joint and separate prediction losses, explicitly quantifying the gain from enlarged effective sample size and the cost of heterogeneity. To attain these rates, we first construct an oracle transfer estimator based on debiasing and truncated normalization. We then develop a computationally efficient procedure that adaptively achieves the minimax rate without prior knowledge of the sparsity levels, using aggregation to control transfer bias and a truncated normalization scheme to preserve transfer gains under the scale-sensitive prediction loss. A data-driven source detection procedure is further proposed to identify potentially informative auxiliary datasets. Our analysis also yields a generalized sin-theta perturbation theorem that removes a commonly imposed comparability condition on leading canonical correlations in the sparse CCA problem.
Related
- Minimax-optimal Scaling Laws in Ridge Regression via Transfer Learning
- Sharp Critical Minimax Laws and No-Learning Thresholds in Continuous-Time Adaptive Control
- Learning Sparse Support under Differential Privacy: Adaptive Algorithms and Minimax Limits
- From Common Information to Canonical Correlation Analysis and Beyond: Bridging Information Theory, Signal Processing, and Machine Learning
- Expected Minimax Regret Rates for Adaptive LQG Control on Fixed Parameter Sets
BibTeX
@misc{fu2026minimax,
title = {{Minimax and Adaptive Transfer Learning for Sparse Canonical Correlation Analysis}},
author = {Bo Fu and Dandan Jiang},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.05735},
eprint = {2610.05735},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.05735},
}
RIS
TY - UNPB
TI - Minimax and Adaptive Transfer Learning for Sparse Canonical Correlation Analysis
AU - Bo Fu
AU - Dandan Jiang
PY - 2026
DA - 2026/10/05
PB - arXiv
DO - 10.48550/arxiv.2610.05735
UR - https://arxiv.org/abs/2610.05735
AB - We develop a transfer learning framework for high-dimensional sparse canonical correlation analysis (CCA) with multiple heterogeneous source datasets. Our goal is to improve estimation of the target canonical coefficient matrices by borrowing information from related sources. We characterize source-target similarity through discrepancies in the joint covariance structure and establish minimax optimal rates under both joint and separate prediction losses, explicitly quantifying the gain from enlarged effective sample size and the cost of heterogeneity. To attain these rates, we first construct an oracle transfer estimator based on debiasing and truncated normalization. We then develop a computationally efficient procedure that adaptively achieves the minimax rate without prior knowledge of the sparsity levels, using aggregation to control transfer bias and a truncated normalization scheme to preserve transfer gains under the scale-sensitive prediction loss. A data-driven source detection procedure is further proposed to identify potentially informative auxiliary datasets. Our analysis also yields a generalized sin-theta perturbation theorem that removes a commonly imposed comparability condition on leading canonical correlations in the sparse CCA problem.
ER -