Skip to main content
U.S. flag

An official website of the United States government

Official websites use .gov
A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS
A lock ( ) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.

Aaron Gilad Kusne (Fed)

Research Scientist, Materials Measurement Science Division, National Institute of Standards and Technology  
Adjunct Associate Professor of Materials Science & Engineering, University of Maryland College Park
Fellow of the American Physical Society

Autonomous Materials Research Systems

Accelerating Discovery & Democratizing Science

Main Projects

Key focus - Machine learning for science: active learning to guide and optimize experiments, incorporating prior scientific knowledge into ML, uncertainty quantification and propagation, trust and interpretability.
Real world successes: discovery of new materials in rare-earth free permanent magnetics, spin-driven thermoelectrics, and the new best-in-class phase change memory material.

Some of our autonomous platforms, i.e., robotic labs:

Examples of our autonomous systems
Credit: Aaron Kusne

Project: Materials Exploration, Discovery, and Optimization
The structure of a material greatly influences its properties. Thus the search for better materials must often include knowledge of the relationship between how a material is made and the resulting structure, i.e. "phase mapping".  
Autonomous phase mapping at the Stanford Linear Accelerator has allowed us to reduce the number of measurement experiments necessary for phase mapping by an order of magnitude. This in turn accelerates materials optimization and discovery, resulting in the first autonomous system to discover a best-in-class material (phase-change memory material) CAMEO: https://www.nature.com/articles/s41467-020-19597-w

Autonomous phase mapping

 

Icons from CAMEO paper available below. Please reference: Kusne, A.G., et al. "On-the-fly closed-loop materials discovery via Bayesian active learning." Nature communications 11.1 (2020): 1-11.
For other versions, please contact me.

CAMEO Icons. Please cite: AG Kusne, https://www.nature.com/articles/s41467-020-19597-w

 

 

 

 


Project: A Scalable Operating System for Laboratories

Autonomous research systems allow scientists to fail smarter, learn faster, and spend less resources in their studies. As these systems grow in number, capability, and complexity, a new challenge arises – how will they work together across large facilities? We developed the operating system "MULTITASK" (MULTI-agent auTonomous fAcilities - a Scalable frameworK), which can 1) manage realistic resource limits such as equipment use, 2) run complex research campaigns via machine learning agents with diverse learning capabilities and goals, and 3) facilitate multi-agent collaborations and teams to parallel benefits of real-world scientist teams. MULTITASK makes possible facility-wide control and simulations, including agent-instrument and agent-agent interactions. Through MULTITASK’s modularity, real-world facilities can come on-line in phases, with simulated instruments gradually replaced by real-world instruments. We hope MULTITASK opens new areas of study in large-scale autonomous and semi-autonomous research campaigns and facilities. Upcoming in Cell Matter. Previous draft on arxiv.

Multitask strageties

 


Project: Autonomous Metrology
We are investigating the use of ML to guide microscopy and other measurement systems to accelerate knowledge capture.

ANDiE (Autonomous Neutron Diffraction Explorer) demonstrates autonomous control over neuron scattering at the NIST Center for Neutron Research and the Oak Ridge National Laboratory High Flux Isotope Reactor. ANDiE achieves a 5x acceleration in determining magnetic structure and transition behavior through novel machine learning with built-in magnetic structure and neutron scattering physics.

https://aip.scitation.org/doi/full/10.1063/5.0082956

 


Project: Low-Cost Autonomous Scientist Kit
We've developed a low cost autonomous scientist kit for Teaching and Developing machine learning.

We have developed the next generation in science education - a kit for building a low-cost autonomous scientist capable of experiment design, execution, and analysis in a closed loop. This kit can be used to teach the next generation workforce in areas such as ML, control systems, measurement science, materials synthesis, decision theory, among others. Industry can also use the kit to develop and evaluate autonomous methodologies. The kit was used during two courses at the University of Maryland to teach undergraduate and graduate students autonomous physical science. The kit was demonstrated for hypothesis design, discovery, and validation.

https://arxiv.org/ftp/arxiv/papers/2204/2204.04187.pdf

 


Project: Autonomous Protein Engineering
The complexity of biological systems is incredible. We are combining ML and robotics to build a greater understanding of protein engineering.  
https://arxiv.org/abs/1911.02106, https://doi.org/10.1101/2020.03.05.979385, https://doi.org/10.1101/2020.07.10.197574

Autonomous Protein Engineering

 


Project: ML for Accelerating Materials Research
We use ML to learn about important materials (e.g. superconductors) and guide research in the lab. https://www.nature.com/articles/s41524-018-0085-8 

ML for Superconductivity

 

 

 

 

 

 

 

 


Project: Bootcamp: Machine Learning for Materials Research
Educating the next generation of physicists and materials scientists.  
For MLMR attendance is approx 30% industry, 10% national labs, and 60% academia.  Over the 7 years of the bootcamp, we have had attendees from a total of 27 countries. We have also run tutorials at MRS, APS, TMS, MLSE, NSF meetings, among others.

https://www.nanocenter.umd.edu/events/mlmr/

 


Project: REMI - REsource for Materials Informatics
A repository for tutorials and code examples covering materials data import/export, pre-processing, and analysis. Search by material system, synthesis / simulation method, measurement method, data type, data analysis type, and more.

https://pages.nist.gov/remi/

 


In the News
ML for Superconductivity: Scientific American: Our ML-driven search for room-temperature superconductors.
CAMEO + Materials Discovery: AAAS Eureka Alert!; Science Daily; Semiconductor Engineering; Phys.orgOptics.orgScience Bulletin; COSMOS MagazineChemical Engineering; UMD News; NIST News


Current Mentees

Haotong Liang
Machine Learning + Materials Science
UMD PhD Student

Chih-Yu Lee
Machine Learning + Materials Science
UMD PhD Student

Felix Adams
Machine Learning + Materials Science
UMD PhD Student

Logan Saar
Machine Learning + Materials Science
UMD Undergraduate Student

Alex Wang
Machine Learning + Materials Science
UMD Undergraduate Student

Dennis Zhao
Machine Learning + Materials Science
UMD Undergraduate Student

 


Past Mentees

Austin McDannald
Past: NIST Postdoc, Machine Learning + Materials Science  
Current: Research Scientist, NIST

Jong Ho Kim
Machine Learning + Materials Science
Visiting Research Scientist, Research Institute of Industrial Science & Technology, Korea

Peter Tonner
Past: NIST NRC Postdoc, Machine Learning + Genetics
Current: Research Scientist, GSK

Brian DeCost
Past: NRC Postdoc, Machine Learning + Materials Science  
Current: Research Scientist, NIST

Graham Antoszewski
Past: UMD Masters Student, Applied Math  
Current: BlackSky

Yuma Iwasaki
Machine Learning + Materials Science
Visiting Research Scientist, NEC Japan

Varshini Salvedurai
Past: Summer High School Student (SHIP)  
Current: CMU CS Undergraduate
 


Open Positions
For openings, please contact me at: aaron(.)kusne(@)nist(.)gov  

NRC Postdoc Postings:



Publication List

See my google scholar page

 


Recently Organized Workshops

2023 8th Annual 5-Day Machine Learning for Materials Research Bootcamp
2022 Advances in Autonomous Materials Research
2022 MRS Biannual Tutorial Day on Data Science
2021 MRS Fall Symposium: Accelerating Experimental Materials Research with Machine Learning
2021 MRS Tutorial Day: Machine Learning for Materials Science and Engineering
2021 Workshop on Autonomous Materials Research
2021 Materials Research Data Alliance Workshop: Education and Workforce Development
2021 LBL Workshop: Workshop on Autonomous Discovery in Science and Engineering
2020 MRS Kavli Workshop: Building a Community for Autonomous Research, Online.
2020 Workshop on Machine Learning Microscopy Data, College Park, MD
2019 Autonomous Systems for Materials Development Workshop, Philadelphia, PA
2019 MLMR Workshop on Autonomous Research, College Park, MD

Awards

Bronze Award - The Bronze Medal Award is the highest recognition awarded by NIST. 

Publications

AI for Materials

Author(s)
Debra Audus, Kamal Choudhary, Brian DeCost, A. Gilad Kusne, Francesca Tavazza, James A. Warren
The application of artificial intelligence (AI) methods to materials re- search and development (MR&D) is poised to radically reshape how materials are

Reproducible Sorbent Materials Foundry for Carbon Capture at Scale

Author(s)
Austin McDannald, Howie Joress, Brian DeCost, Avery Baumann, A. Gilad Kusne, Kamal Choudhary, Taner N. Yildirim, Daniel Siderius, Winnie Wong-Ng, Andrew J. Allen, Christopher Stafford, Diana Ortiz-Montalvo
We envision an autonomous sorbent materials foundry (SMF) for rapidly evaluating materials for direct air capture of carbon dioxide ( CO2), specifically

Patents (2018-Present)

Novel Nanocomposite Phase-Change Memory Materials And Design Of The Same

NIST Inventors
Ichiro Takeuchi and Aaron Gilad Kusne
Provided herein are novel materials, such as novel phase-change memory materials providing superior characteristics, and methods of discovering/selecting such novel materials via machine learning, such as Bayesian active learning. An exemplary material provided by the inventive concept is the

Novel Nanocomposite Phase-Change Memory Materials And Design And Selection Of The Same

NIST Inventors
Ichiro Takeuchi and Aaron Gilad Kusne
Provided herein are novel materials, such as novel phase-change memory materials providing superior characteristics, and methods of discovering/selecting such novel materials via machine learning, such as Bayesian active learning. An exemplary material provided by the inventive concept is the
Created April 7, 2019, Updated September 25, 2024