Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

SMRS: advocating a unified reporting standard for surrogate models in the artificial intelligence era

Elizaveta Semenova, Alisa Sheinkman, Timothy James Hitge, Siobhan Mackenzie Hall, Jon Cockayne

TL;DR

Surrogate models are increasingly used across disciplines but lack standardized reporting, hindering reproducibility and cross-domain transfer. The authors propose SMRS, a lightweight, modular framework that documents data provenance, sampling design, model assumptions, learning methods, evaluation, and uncertainty quantification in a model-agnostic format. Through a review of 17 studies and several case studies, SMRS highlights current reporting fragmentation and demonstrates how a unified schema can improve transparency, benchmarking, and cross-domain reuse. By enabling concise, auditable reporting and practical adoption pathways, SMRS aims to accelerate trustworthy, interoperable surrogate modelling in the AI era.

Abstract

Surrogate models are widely used to approximate complex systems across science and engineering to reduce computational costs. Despite their widespread adoption, the field lacks standardisation across key stages of the modelling pipeline, including data sampling, model selection, evaluation, and downstream analysis. This fragmentation limits reproducibility and cross-domain utility -- a challenge further exacerbated by the rapid proliferation of AI-driven surrogate models. We argue for the urgent need to establish a structured reporting standard, the Surrogate Model Reporting Standard (SMRS), that systematically captures essential design and evaluation choices while remaining agnostic to implementation specifics. By promoting a standardised yet flexible framework, we aim to improve the reliability of surrogate modelling, foster interdisciplinary knowledge transfer, and, as a result, accelerate scientific progress in the AI era.

SMRS: advocating a unified reporting standard for surrogate models in the artificial intelligence era

TL;DR

Surrogate models are increasingly used across disciplines but lack standardized reporting, hindering reproducibility and cross-domain transfer. The authors propose SMRS, a lightweight, modular framework that documents data provenance, sampling design, model assumptions, learning methods, evaluation, and uncertainty quantification in a model-agnostic format. Through a review of 17 studies and several case studies, SMRS highlights current reporting fragmentation and demonstrates how a unified schema can improve transparency, benchmarking, and cross-domain reuse. By enabling concise, auditable reporting and practical adoption pathways, SMRS aims to accelerate trustworthy, interoperable surrogate modelling in the AI era.

Abstract

Surrogate models are widely used to approximate complex systems across science and engineering to reduce computational costs. Despite their widespread adoption, the field lacks standardisation across key stages of the modelling pipeline, including data sampling, model selection, evaluation, and downstream analysis. This fragmentation limits reproducibility and cross-domain utility -- a challenge further exacerbated by the rapid proliferation of AI-driven surrogate models. We argue for the urgent need to establish a structured reporting standard, the Surrogate Model Reporting Standard (SMRS), that systematically captures essential design and evaluation choices while remaining agnostic to implementation specifics. By promoting a standardised yet flexible framework, we aim to improve the reliability of surrogate modelling, foster interdisciplinary knowledge transfer, and, as a result, accelerate scientific progress in the AI era.

Paper Structure

This paper contains 25 sections, 1 figure.

Table of Contents

  1. Introduction
  2. Proposed reporting schema for surrogate models
  3. Do you really need a surrogate?
  4. Data collection and generation
  5. Two modes of input data.
  6. Sampling design.
  7. Standardised data formats.
  8. Model class selection
  9. Learning methods
  10. Evaluation metrics and benchmarks
  11. Diagnostic evaluation: distributional similarity.
  12. Diagnostic evaluation: calibration.
  13. Diagnostic evaluation: estimated generalisation performance of the model.
  14. Task-based evaluation.
  15. Benchmarking.
  16. ...and 10 more sections

Figures (1)

  • Figure 1: An Overview of the proposed Surrogate Model Reporting Standard (SMRS)