What does this research mean for the field?

An active learning approach that leverages the joint posterior distribution of Gaussian process gradients creates acquisition functions that improve the accuracy and robustness of global sensitivity analysis compared to existing derivative-based criteria. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

What question did this study set out to answer?

Develop an active learning framework leveraging Gaussian process gradient distributions to optimize global sensitivity analysis under strictly limited evaluation budgets.

September 5, 2026ESAIM Probability and StatisticsOpen Access

Gradient-based active learning with Gaussian processes for global sensitivity analysis.

View Full Paper

Ask AI

Bookmark

Develop an active learning framework leveraging Gaussian process gradient distributions to optimize global sensitivity analysis under strictly limited evaluation budgets.
Constructed acquisition functions using the joint posterior distribution of Gaussian process gradients to capture correlations between partial derivatives.
Integrated derivative-based global sensitivity measures and Sobol' indices into an adaptive experimental design strategy.
Validated the approach against state-of-the-art benchmarks and applied it to an environmental simulator modeling pesticide transfer.
Demonstrated superior sensitivity analysis accuracy and robustness compared to standard derivative-based acquisition criteria across benchmark functions.
Successfully identified key influential inputs in a computationally intensive real-world environmental pesticide transfer model with minimal evaluations.

Discussion

No takes yet. Share an insight, caveat, or question.

Implication

Simulation study demonstrates improved sensitivity analysis accuracy via gradient-based active learning in complex numerical models, indicating enhanced efficiency under constrained budgets.

Cite This Study

Lambert et al. (2026) studied this question.

synapsesocial.com/papers/6a9bd3726b95aff0620eaae0 https://doi.org/10.1051/ps/2026017

View Full Paper

Ask AI

Bookmark

Implication

Simulation study demonstrates improved sensitivity analysis accuracy via gradient-based active learning in complex numerical models, indicating enhanced efficiency under constrained budgets.

Gradient-based active learning with Gaussian processes for global sensitivity analysis.

Key Points

Discussion

Implication

Cite This Study

Gradient-based active learning with Gaussian processes for global sensitivity analysis.

Key Points

Discussion

Implication

Cite This Study

Also Consider

Also Consider