Optimizing Autonomous and Human-Assisted Experimentation in Materials Development

Abstract

Advances in machine intelligence have the potential to increase the capabilities of autonomous research systems and human-machine research systems. In this interdisciplinary proposal, we explore the application of these advances for accelerating materials development in both systems. The proposed 3-year project extends our on-going collaboration with Dr. Maruyama’s materials science team in the Air Force Research Laboratory, and focuses on three complementary lines of closed-loop research: (1) Multi-objective optimization algorithms will be developed to automate the synthesis of carbon nanotubes that are customized to given specifications; (2) Human-guided Bayesian optimization algorithms will be explored in additive manufacturing (3D printing) where evaluation of product quality by human experts may be superior to sensor measurements; (3) Automated machine learning approaches will be applied to the autonomous research systems in (1) and (2) to further improve performance while saving time and effort for fine-tuning the systems. Results of this work should also advance discovery in other areas of materials science.

Document Details

Document Type
DoD Grant Award
Publication Date
Jan 21, 2022
Source ID
FA95502110176XX0

Entities

People

  • Mark Pitt

Organizations

  • Air Force Office of Scientific Research
  • Ohio State University
  • United States Air Force

Tags

Readers

  • Distributed Systems and Data Platform Development
  • Neural Network Machine Learning.
  • Research Science/Academic Research

Technology Areas

  • AI & ML
  • AI & ML - DoD AI Strategy