# Local transfer learning Gaussian process modeling, with applications to surrogate modeling of expensive computer simulators

arXiv: https://arxiv.org/abs/2410.12690 · HTML: https://mlsanity.com/paper/2410.12690/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 17 of its 67 references that are in the map.
- First posted on arXiv 16 Oct 2024.

## Builds on (most cited first, 17 of 17)

- [A Comprehensive Survey on Transfer Learning](https://mlsanity.com/paper/1911.02685/) (2019)
- [Automatic differentiation in machine learning: a survey](https://mlsanity.com/paper/1502.05767/) (2015)
- [Deep Neural Networks as Gaussian Processes](https://mlsanity.com/paper/1711.00165/) (2017)
- [AFEC: Active Forgetting of Negative Transfer in Continual Learning](https://mlsanity.com/paper/2110.12187/) (2021)
- [Multi-fidelity Bayesian Neural Networks: Algorithms and Applications](https://mlsanity.com/paper/2012.13294/) (2020)
- [Active Learning for Deep Gaussian Process Surrogates](https://mlsanity.com/paper/2012.08015/) (2020)
- [A hierarchical expected improvement method for Bayesian optimization](https://mlsanity.com/paper/1911.07285/) (2019)
- [On Negative Transfer and Structure of Latent Functions in Multi-output Gaussian Processes](https://mlsanity.com/paper/2004.02382/) (2020)
- [A graphical multi-fidelity Gaussian process model, with application to emulation of heavy-ion collisions](https://mlsanity.com/paper/2108.00306/) (2021)
- [Additive Multi-Index Gaussian process modeling, with application to multi-physics surrogate modeling of the quark-gluon plasma](https://mlsanity.com/paper/2306.07299/) (2023)
- [Transfer learning driven design optimization for inertial confinement fusion](https://mlsanity.com/paper/2205.13519/) (2022)
- [Adaptive design for Gaussian process regression under censoring](https://mlsanity.com/paper/1910.05452/) (2019)
- [Conglomerate Multi-Fidelity Gaussian Process Modeling, with Application to Heavy-Ion Collisions](https://mlsanity.com/paper/2209.13748/) (2022)
- [Hierarchical shrinkage Gaussian processes: applications to computer code emulation and dynamical system recovery](https://mlsanity.com/paper/2302.00755/) (2023)
- [Regularized Multi-output Gaussian Convolution Process with Domain Adaptation](https://mlsanity.com/paper/2409.02778/) (2024)
- [Targeted Variance Reduction: Robust Bayesian Optimization of Black-Box Simulators with Noise Parameters](https://mlsanity.com/paper/2403.03816/) (2024)
- [Bayesian analysis of multifidelity computer models with local features and non-nested experimental designs: Application to the WRF model](https://mlsanity.com/paper/1910.08063/) (2019)

## Built on by (most cited first, 1 of 1)

- [Multifidelity Simulation-based Inference for Computationally Expensive Simulators](https://mlsanity.com/paper/2502.08416/) (2025)

## Abstract (arXiv)

A critical bottleneck for scientific progress is the costly nature of computer simulations for complex systems. Surrogate models provide an appealing solution: such models are trained on simulator evaluations, then used to emulate and quantify uncertainty on the expensive simulator at unexplored inputs. In many applications, one often has available data on related systems. For example, in designing a new jet turbine, there may be existing studies on turbines with similar configurations. A key question is how information from such ``source'' systems can be transferred for effective surrogate training on the ``target'' system of interest. We thus propose a new LOcal transfer Learning Gaussian Process (LOL-GP) model, which leverages a carefully-designed Gaussian process to transfer such information for surrogate modeling. The key novelty of the LOL-GP is a latent regularization model, which identifies regions where transfer should be performed and regions where it should be avoided. Such a ``local transfer'' property is present in many scientific systems: at certain parameters, systems may behave similarly and thus transfer is beneficial; at other parameters, they may behave differently and thus transfer is detrimental. By accounting for local transfer, the LOL-GP can temper the risk of ``negative transfer'', i.e., the risk of worsening predictive performance from information transfer. We derive a Gibbs sampling algorithm for efficient posterior predictive sampling on the LOL-GP, for both the multi-source and multi-fidelity transfer settings. We then show, via a suite of numerical experiments and an application for jet turbine design, the improved surrogate performance of the LOL-GP over existing methods.