Abstract:Multifidelity surrogate modelling combines data of varying accuracy and cost from different sources. It strategically uses low-fidelity models for rapid evaluations, saving computational resources, and high-fidelity models for detailed refinement. It improves decision-making by addressing uncertainties and surpassing the limits of single-fidelity models, which either oversimplify or are computationally intensive. Blending high-fidelity data for detailed responses with frequent low-fidelity data for quick approximations facilitates design optimisation in various domains.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: how to effectively fuse multifidelity data to construct more accurate and efficient surrogate models, thereby improving the accuracy of decision - making and reducing computational costs in design optimization and other applications. ### Specific problems include: 1. **Data fusion challenges**: - How to select an appropriate fidelity level. - How to develop effective data fusion methods, especially when dealing with data of different precisions and computational costs. 2. **Limitations of single - fidelity models**: - Single - fidelity models are either overly simplified in terms of actual physical phenomena or too computationally expensive to be used effectively. - Multifidelity surrogate models can overcome these limitations by combining the rapid evaluation of low - fidelity models and the detailed corrections of high - fidelity models. 3. **Deficiencies of existing methods**: - Although progress has been made in interpolation, regression, enhanced sampling, error estimation, variable fidelity, and data fusion techniques, challenges still exist in selecting fidelity levels and developing efficient data fusion methods. 4. **Proposing a new fusion method**: - This research proposes a new fusion method to achieve the fusion of multifidelity data by constructing regression surfaces using only gradient information (gradient - only surrogates). - This method aims to avoid the confusion caused by complex examples and focuses on basic example problems to clearly demonstrate the effectiveness of the new method. ### Research background and significance: - **Application areas**: Multifidelity surrogate models are widely used in fields such as aerospace, automotive design, materials science, financial risk management, and environmental modeling. - **Innovation points**: By giving priority to using gradient information rather than function values, it challenges the traditional concept that more data does not necessarily lead to better results. Instead, carefully selecting high - quality data can significantly improve model performance. - **Future directions**: It emphasizes the importance of data quality and relevance, providing new ideas for future multifidelity modeling and optimization strategies. In conclusion, this paper aims to improve the existing surrogate model construction techniques by proposing a gradient - based multifidelity data fusion method, thereby achieving more efficient design optimization in various engineering and scientific applications.

Multifidelity Surrogate Models: A New Data Fusion Perspective

A novel multi-fidelity surrogate modeling method for non-hierarchical data fusion

A multi-fidelity surrogate modeling method in the presence of non-hierarchical low-fidelity data

Multifidelity linear regression for scientific machine learning from scarce data

A novel multi-fidelity surrogate modeling framework integrated with sequential sampling criterion for non-hierarchical data

A survey on multi-fidelity surrogates for simulators with functional outputs: unified framework and benchmark

Multi-fidelity reduced-order surrogate modelling

Practical multi-fidelity machine learning: fusion of deterministic and Bayesian models

A comprehensive framework for multi-fidelity surrogate modeling with noisy data: a gray-box perspective

Methodology And Challenges Of Surrogate Modelling Methods For Multi-Fidelity Expensive Black-Box Problems

Enhanced multi-fidelity modelling for digital twin and uncertainty quantification

Characterising harmful data sources when constructing multi-fidelity surrogate models

A Single-Fidelity Surrogate Modeling Method Based on Nonlinearity Integrated Multi-Fidelity Surrogate

Multi-fidelity information fusion with hierarchical surrogate guided by feature mapping

Hyperparameter-optimized Multi-fidelity Deep Neural Network model associated with Subset Simulation for structural reliability analysis

Adaptive Multifidelity Data Assimilation for Nonlinear Subsurface Flow Problems

A comprehensive multi-fidelity surrogate framework based on Gaussian process for datasets with heterogeneous responses

Multi-Fidelity Residual Neural Processes for Scalable Surrogate Modeling

Multi-fidelity Gaussian process surrogate modeling for regression problems in physics

Assessment of Multifidelity Surrogate Approaches for Expedient Loads Prediction in High-Speed Flows

Use of Multifidelity Training Data and Transfer Learning for Efficient Construction of Subsurface Flow Surrogate Models