Materials have historically been developed by trial and error in research and development laboratories. This process has largely happened in isolation from product design and manufacturing. As a result, it has typically taken a long time for new materials to impact the marketplace. Moreover, new discoveries of potential materials are often lost or revisited/rediscovered due to inefficiencies in archival and use of research results. In view of the explosive growth of data storage, computational modeling and simulation, data sciences, and informatics, it is increasingly possible to envision a future where materials researchers, suppliers, product designers, and manufacturers work more closely and exchange information to reduce the time for materials development and insertion. Digital data is the currency that will fuel materials research and development collaborations.
The Georgia Tech Institute for Materials is supporting development of a web-based environment hosting open source software tools that add value by enhancing collaboration and distributed access to experimental and computational data. Moreover, workflows of collaborative efforts can be tracked. Called MATIN, this digital environment is being developed in alignment with the philosophy of the Materials Genome Initiative (MGI). Primary elements of GT’s unique and innovative approach include (i) development and deployment of customized data engineering and science solutions to address the central challenge of scale-bridging (also referred to as “tyranny of scales”) for seamless exchange of high value, reliable, information (i.e., knowledge) in both directions between constituent hierarchical scales; (ii) utilization of advanced concepts in statistical sciences for rigorous quantification of the uncertainty associated with the information exchange that occurs during scale-bridging and providing objective decision support for multiscale materials design exploration efforts; (iii) world-class expertise in multiscale modeling that cuts across the gamut of hierarchical structure scales encountered in virtually all materials classes of interest to advanced technologies in transportation, health, energy, and defense industries; (iv) unique high throughput strategies that harmoniously blend together computational, experimental, and data science tools to facilitate rapid and efficient exploration of unimaginably large design spaces in hierarchical materials systems; and (v) open source, open access, e-collaboration (online) environments that are specifically designed to dramatically enhance the productivity of cross-disciplinary project teams by facilitating easy sharing and exchange of information, knowledge, and expertise between geographically and organizationally distributed team members spread across diverse stakeholder groups.
The Georgia Tech MATIN e-collaborative platform seeks to support high throughput materials discovery, development and deployment by considering an open source software approach, combined with in-house tools that include:
- Advanced Statistics
- Uncertainty Quantification
- Objective Decision Support Systems
- Ontologies in digital representation of hierarchical material structures
- Knowledge Mining using machine learning
- Computer Vision
- Data Management
- Digital Capture, Curation, and Dissemination
- Support for Inductive, Top Down Materials Design Exploration
Many of these tools focus on process-structure-property relations at the mesoscale, which is key to unlocking the potential of many material systems with hierarchical structure. Examples of MATIN applications include:
- Multivariate Polynomial Regression
- Spatial Statistics
- Project Pages for Research e-Collaboration
The FLAMEL program is an important joint initiative between IMat and Computational Science and Engineering within the College of Computing aimed at preparing the future workforce in materials data sciences and informatics. This NSF Integrated Graduate Education and Research Traineeship Program (IGERT) is anticipated to support the research of over 20 doctoral students working at the interface of materials research and data sciences from 2014 – 2019.
Through its leadership role in hosting regional and national workshops and planning activities for the MGI, IMat is also pursuing materials data sciences and informatics on the national scale as an element of the national materials accelerator network