Technology Category
Machine Learning: Transfer, Domain Adaptation, Multi-Task LearningIntelligent Robots: Multimodal Perception & Sensor FusionSearch and Optimization: Sampling/Simulation-based Search
Multi-task Modeling for Engineering Applications with Sparse Data
🤖 AI Summary
This study addresses the challenge of scarce high-fidelity data amid abundant low-fidelity observations in engineering systems by proposing a multitask Gaussian process (MTGP) framework that explicitly models cross-task correlations among output variables and multiple fidelity levels. The approach innovatively integrates multi-source, multi-fidelity information with a structured dependency among multiple outputs, enabling efficient and robust collaborative prediction even under extreme data sparsity. Experimental validation on three representative engineering scenarios—the Forrester function, three-dimensional ellipsoidal cavity modeling, and friction stir welding—demonstrates that the proposed method significantly enhances predictive accuracy, thereby offering effective support for intelligent decision-making in high-cost engineering environments.
Technology Category
Application Category
📝 Abstract
Modern engineering and scientific workflows often require simultaneous predictions across related tasks and fidelity levels, where high-fidelity data is scarce and expensive, while low-fidelity data is more abundant. This paper introduces an Multi-Task Gaussian Processes (MTGP) framework tailored for engineering systems characterized by multi-source, multi-fidelity data, addressing challenges of data sparsity and varying task correlations. The proposed framework leverages inter-task relationships across outputs and fidelity levels to improve predictive performance and reduce computational costs. The framework is validated across three representative scenarios: Forrester function benchmark, 3D ellipsoidal void modeling, and friction-stir welding. By quantifying and leveraging inter-task relationships, the proposed MTGP framework offers a robust and scalable solution for predictive modeling in domains with significant computational and experimental costs, supporting informed decision-making and efficient resource utilization.
Problem
Research questions and friction points this paper is trying to address.
multi-task learning
multi-fidelity data
data sparsity
engineering applications
predictive modeling
Innovation
Methods, ideas, or system contributions that make the work stand out.
Multi-Task Gaussian Processes
Multi-fidelity modeling
Sparse data
Inter-task correlation
Engineering applications
🔎 Similar Papers
No similar papers found.
