Compressed Active Subspaces for Scalable Bayesian Inference
Abstract
Active subspace methods provide a framework for quantifying predictive uncertainty in high-dimensional models by identifying and performing inference along parameter directions that have the greatest influence on the model output. However, the construction of active subspaces requires storing many full-dimensional model gradients, which becomes prohibitive as model size increases. We address this limitation by proposing Compressed Active Subspaces (CAS), a scalable approach that first maps the model parameters to a compressed space using a structured isometric embedding and then constructs the active subspace within this reduced parameterization. Our approach substantially reduces the memory required for active subspace construction and enables Bayesian inference for large models where standard active subspace methods become impractical. We demonstrate the scalability of CAS on neural networks of increasing size while maintaining predictive performance and robust uncertainty estimates.
Related
- Scalable Bayesian Inference for Time Series via Divide and Conquer
- Subsampled Pseudo-posteriors for Scalable Bayesian Moment-condition Inference
- Scalable Bayesian inference for time-discretised Markov jump processes via exchangeable particle Gibbs
- Cheap and Powerful Tests for Supervised Subspaces: Per-Component Inference for PLS
- Balanced Active Subspaces for Optimal Control via Trajectory Optimization
BibTeX
@misc{flynn2026compressed,
title = {{Compressed Active Subspaces for Scalable Bayesian Inference}},
author = {Thomas Flynn and Sanket Jantre and Byung-Jun Yoon and Kibaek Kim},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.19539},
eprint = {2609.19539},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.19539},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Compressed Active Subspaces for Scalable Bayesian Inference
AU - Thomas Flynn
AU - Sanket Jantre
AU - Byung-Jun Yoon
AU - Kibaek Kim
PY - 2026
DA - 2026/09/17
PB - arXiv
DO - 10.48550/arxiv.2609.19539
UR - https://arxiv.org/abs/2609.19539
AB - Active subspace methods provide a framework for quantifying predictive uncertainty in high-dimensional models by identifying and performing inference along parameter directions that have the greatest influence on the model output. However, the construction of active subspaces requires storing many full-dimensional model gradients, which becomes prohibitive as model size increases. We address this limitation by proposing Compressed Active Subspaces (CAS), a scalable approach that first maps the model parameters to a compressed space using a structured isometric embedding and then constructs the active subspace within this reduced parameterization. Our approach substantially reduces the memory required for active subspace construction and enables Bayesian inference for large models where standard active subspace methods become impractical. We demonstrate the scalability of CAS on neural networks of increasing size while maintaining predictive performance and robust uncertainty estimates.
ER -