Active learning framework for optimized formulation design
This solution integrates machine learning with optimization algorithms to streamline the formulation process, focusing on efficient identification of optimal conditions for multiple properties through minimal testing.
Overview
This active learning framework is designed to optimize formulation processes by integrating machine learning with advanced optimization algorithms like genetic algorithms and Bayesian optimization. The technology aims to efficiently search for the best formulation configurations by predicting and refining conditions with minimal testing requirements. This approach breaks the trade-off between thermodynamic and dynamic properties in biocondensates, offering a robust solution for achieving multiple desired properties simultaneously.
Technical specifications
Key features:
- Utilizes Gaussian process regressor for data point identification, enhancing accuracy with minimal testing.
- Employs genetic algorithms and Bayesian optimization for robust optimization of multiple objectives.
- Includes reinforced learning for dynamic control of formulations, adjusting outputs as they deviate from targets.
- Data preprocessing to ensure high-quality datasets, including outlier removal and feature correlation analysis.
- Iterative model training focusing on refining predictions with minimal testing benchmarks.
Technology readiness level
This solution is currently at Technology Readiness Level 4, indicating that it has been validated in a lab environment and is progressing towards further development and validation.
About Louisiana State University
Louisiana State University is a flagship public land-grant university in Baton Rouge, serving a large student body and a comprehensive research enterprise. A research and technology park provides incubator space, corporate labs, and direct access to campus expertise. A statewide extension network and field sites connect companies to real-world testing, and system-wide health sciences partners support clinical collaboration. Gulf Coast location offers access to suppliers, infrastructure, and demonstration sites. Research is backed by federal funding from NSF, NIH, DOE, USDA, and DoD, and a technology transfer office streamlines IP and startup formation.
delivered to your inbox.