HAL rapidly determines the relationships between inputs and outputs and places new design points in dynamic regions and catalogs findings. It empowers users with interactive visualizations of high-dimensional data to identify driving factors and areas of interest for further exploration.

Features

• State-of-the-art sampling techniques for deep exploration of high-dimensional models
• Automatically compares built-in, cutting-edge machine learning techniques to find the best meta-model for the job
• Produces detailed, automated V&V support reports, enabling full understanding of the harnessed model
• Offers rapid comparisons between baseline and revised models, mapping identical input designs and providing an automated comparison report
• Automates the distribution of modeling computation to include leveraging Department of Defense (DoD) High-Performance Computing (HPC)
• Provides data visualization and exploration to identify and visualize patterns, anomalies and trends across observed data and meta-models

Benefits

• HAL is model agnostic and is architected to easily integrate the latest libraries for experimental design/sampling, machine learning and analysis
• Accelerates analytical workflows and model development
• Alleviates the analytic burden of developing experimental designs
• Promotes best practices
• Improves workforce skill