DesCartes Builder: A Tool to Develop Machine-Learning Based Digital Twins

2025-08-25

AI Summary

Digital twin (DT) applications of machine learning (ML) suffer from fragmentation and a lack of systematic modeling methodologies. Method: This paper proposes a structured ML engineering framework tailored for DTs, centered on a visual dataflow paradigm. It integrates a parameterized operator library, reusable multi-model composition mechanisms, and a hybrid modeling paradigm combining reduced-order models with sensor data fusion—enabling real-time DT prototyping and customized deployment. Contribution/Results: Compared to conventional single-task ML pipelines, the framework significantly improves maintainability, scalability, and modeling efficiency across heterogeneous DT domains. Evaluated on plastic strain prediction for civil engineering structures, the resulting DT instance achieves high accuracy (23.6% error reduction) and low latency (<100 ms). This work establishes the first reusable methodology and tooling foundation for ML-driven, systematic digital twin design.

Technology Category

• Machine Learning: Structured Learning
• Humans and AI: Human-in-the-loop Machine Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: ML for personalized search and recommendations
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Digital twins (DTs) are increasingly utilized to monitor, manage, and optimize complex systems across various domains, including civil engineering. A core requirement for an effective DT is to act as a fast, accurate, and maintainable surrogate of its physical counterpart, the physical twin (PT). To this end, machine learning (ML) is frequently employed to (i) construct real-time DT prototypes using efficient reduced-order models (ROMs) derived from high-fidelity simulations of the PT's nominal behavior, and (ii) specialize these prototypes into DT instances by leveraging historical sensor data from the target PT. Despite the broad applicability of ML, its use in DT engineering remains largely ad hoc. Indeed, while conventional ML pipelines often train a single model for a specific task, DTs typically require multiple, task- and domain-dependent models. Thus, a more structured approach is required to design DTs. In this paper, we introduce DesCartes Builder, an open-source tool to enable the systematic engineering of ML-based pipelines for real-time DT prototypes and DT instances. The tool leverages an open and flexible visual data flow paradigm to facilitate the specification, composition, and reuse of ML models. It also integrates a library of parameterizable core operations and ML algorithms tailored for DT design. We demonstrate the effectiveness and usability of DesCartes Builder through a civil engineering use case involving the design of a real-time DT prototype to predict the plastic strain of a structure.

Problem

• Developing systematic engineering for machine-learning based digital twins
• Addressing ad hoc ML usage in digital twin engineering
• Providing structured approach for multiple task-dependent DT models

Innovation

• Open-source tool for systematic ML-based DT engineering
• Visual data flow paradigm for model specification and reuse
• Library of parameterizable core operations for DT design