Adaptive multi-resolution Gaussian processes: Scalable exact inference with naturally data-sparse covariance matrices

2026-09-24

AI Summary

This study addresses the bottleneck in Gaussian processes where computational efficiency and model fidelity are difficult to balance on large-scale data. To overcome this, we propose an adaptive multi-resolution Gaussian process framework. Specifically, the method designs multi-resolution basis functions anchored on samples to eliminate the reliance on inducing points, constructs naturally sparse covariance matrices via support domain contraction, and integrates a sparse inverse Cholesky algorithm to enable scalable exact inference. The proposed approach reduces training complexity to O(n log²n) and prediction complexity to O(log^d n), achieving high-fidelity large-scale regression without requiring auxiliary points.

Abstract

Gaussian processes constitute a cornerstone of probabilistic machine learning, yet scaling them to large datasets typically forces a trade-off between computational efficiency and model fidelity. This work bridges this gap by presenting an adaptive multi-resolution Gaussian process framework that is both scalable and exact. Our key innovation is constructing a naturally data-sparse covariance matrix with adaptive multi-resolution basis functions. These basis functions are directly anchored to samples, eliminating the need for auxiliary points. By shrinking the support domains of multi-resolution basis, the matrix block sizes are limited, guaranteeing sparsity. The inverse of the data-sparse covariance matrix is computed exactly and efficiently via the sparse Cholesky inverse algorithm. To further improve predictive uncertainties, we construct an augmented basis function. Theoretical analysis and numerical experiments demonstrate that our model achieves exact inference with $\mathcal{O}(n \log^2 n)$ training cost and $\mathcal{O}(\log^d n)$ prediction cost, establishing a principled framework for scalable and high-fidelity Gaussian process regression.

Problem

• Gaussian processes
• scalability
• exact inference
• large datasets
• computational efficiency

Innovation

• Adaptive multi-resolution Gaussian processes
• Data-sparse covariance matrix
• Sparse Cholesky inverse
• Scalable exact inference
• Augmented basis function