ML_LEARN_ERSATZ

ML_LEARN_ERSATZ Worker

Overview

A new worker, ML_LEARN_ERSATZ, has been added to Workflows. This worker enables training of machine learning models to generate Ersatz (surrogate) models based on input data.

Description

The ML_LEARN_ERSATZ worker is designed to learn relationships between input parameters and target responses, creating a surrogate model that can approximate complex simulations or experimental results. This helps reduce computation time by replacing expensive evaluations with fast predictions.

ML_LEARN_ERSATZ — Material Metadata Input

Overview

The ML_LEARN_ERSATZ workflow now supports a new input named Material Metadata. This input allows users to provide additional dataset-based material information directly into the workflow for enhanced model configuration and processing.

Feature Details

• Added a new workflow input: Material Metadata
• Input type: Dataset
• Enables structured material-related metadata to be passed into the workflow
• Supports improved material mapping and metadata-driven processing within ML_LEARN_ERSATZ