Case for a unified surrogate modelling framework in the age of AI

2025-02-10

AI Summary

Surrogate modeling in natural sciences, engineering, and machine learning suffers from nonstandardized practices in data sampling, model selection, evaluation metrics, and downstream task validation—undermining reproducibility and cross-domain transferability. Method: This work introduces the first comprehensive, AI-era surrogate modeling lifecycle standardization framework. It systematically integrates active learning, multi-fidelity sampling, interpretability-aware evaluation, and task-oriented validation, while supporting both classical statistical surrogates and emerging AI-native surrogates (e.g., neural operators, diffusion-based surrogates). Contribution/Results: The framework balances generality with domain-specific adaptability, establishing foundational principles and implementation guidelines. It enhances surrogate reliability, facilitates interdisciplinary knowledge integration, and provides a consensus-driven foundation for advancing scientific computing paradigms.

Technology Category

• Machine Learning: Active Learning
• Search and Optimization: Sampling/Simulation-based Search
• Reasoning under Uncertainty: Stochastic Optimization

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems

Abstract

Surrogate models are widely used in natural sciences, engineering, and machine learning to approximate complex systems and reduce computational costs. However, the current landscape lacks standardisation across key stages of the pipeline, including data collection, sampling design, model class selection, evaluation metrics, and downstream task performance analysis. This fragmentation limits reproducibility, reliability, and cross-domain applicability. The issue has only been exacerbated by the AI revolution and a new suite of surrogate model classes that it offers. In this position paper, we argue for the urgent need for a unified framework to guide the development and evaluation of surrogate models. We outline essential steps for constructing a comprehensive pipeline and discuss alternative perspectives, such as the benefits of domain-specific frameworks. By advocating for a standardised approach, this paper seeks to improve the reliability of surrogate modelling, foster cross-disciplinary knowledge transfer, and, as a result, accelerate scientific progress.

Problem

• Standardize surrogate model pipeline stages
• Enhance reproducibility and cross-domain applicability
• Advocate unified framework for model reliability

Innovation

• Unified surrogate modelling framework
• Standardisation across pipeline stages
• Cross-domain applicability improvement