An official website of the United States government
A first foundational assessment is provided for disaster debris reconnaissance that includes identifying tools and techniques for reconnaissance activities, identifying challenges in field reconnaissance, and identifying and developing preliminary guidelines and standards based on advancements from a workshop held in 2022. In this workshop, reconnaissance activities were analyzed in twofold: in relation to post-disaster debris and waste materials and in relation to waste management infrastructure. A four-phase timeline was included to capture the full lifecycle of management activities ranging from collection to temporary storage to final management route: pre-disaster or pre-reconnaissance, post-disaster response (days/weeks), short-term recovery (weeks/months), and long-term recovery (months/years). For successful reconnaissance, objectives of field activities and data collection needs; data types and metrics; and measurement and determination methods need to be identified. A reconnaissance framework, represented using a 3x2x2x4 matrix, is proposed to incorporate data attributes (tools, challenges, guides), reconnaissance attributes (debris, infrastructure; factors, actions), and time attributes (pre-event, response, short-term, long-term). This framework supports field reconnaissance missions and protocols that are longitudinally based and focused on post-disaster waste material and infrastructure metrics that advance sustainable materials management practices. To properly frame and develop effective reconnaissance activities, actions for all data attributes (tools, challenges, guides) are proposed to integrate sustainability and resilience considerations. While existing metrics, tools, methods, standards, and protocols can be adapted for sustainable post-disaster materials management reconnaissance, development of new approaches are needed for addressing unique aspects of disaster debris management.
more »
« less
The laboratory of Babel: highlighting community needs for integrated materials data management
Automated experimentation methods are unlocking a new data-rich research paradigm in materials science that promises to accelerate the pace of materials discovery. However, if our data management practices do not keep pace with progress in automation, this revolution threatens to drown us in unusable data. In this perspective, we highlight the need to update data management practices to track, organize, process, and share data collected from laboratories with deeply integrated automation equipment. We argue that a holistic approach to data management that integrates multiple scales (experiment, group and community scales) is needed. We propose a vision for what this integrated data future could look like and compare existing work against this vision to find gaps in currently available data management tools. To realize this vision, we believe that development of standard protocols for communicating with equipment and data sharing, the development of new open-source software tools for managing data in research groups, and leadership and direction from funding agencies and other organizations are needed.
more »
« less
- PAR ID:
- 10456015
- Date Published:
- Journal Name:
- Digital Discovery
- Volume:
- 2
- Issue:
- 3
- ISSN:
- 2635-098X
- Page Range / eLocation ID:
- 544 to 556
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Large Language Models (LLMs) are reshaping many aspects of materials science and chemistry research, enabling advances in molecular property prediction, materials design, scientific automation, knowledge extraction, and more. Recent developments demonstrate that the latest class of models are able to integrate structured and unstructured data, assist in hypothesis generation, and streamline research workflows. To explore the frontier of LLM capabilities across the research lifecycle, we review applications of LLMs through 34 total projects developed during the second annual Large Language Model Hackathon for Applications in Materials Science and Chemistry, a global hybrid event. These projects spanned seven key research areas: (1) molecular and material property prediction, (2) molecular and material design, (3) automation and novel interfaces, (4) scientific communication and education, (5) research data management and automation, (6) hypothesis generation and evaluation, and (7) knowledge extraction and reasoning from the scientific literature. Collectively, these applications illustrate how LLMs serve as versatile predictive models, platforms for rapid prototyping of domain-specific tools, and much more. In particular, improvements in both open source and proprietary LLM performance through the addition of reasoning, additional training data, and new techniques have expanded effectiveness, particularly in low-data environments and interdisciplinary research. As LLMs continue to improve, their integration into scientific workflows presents both new opportunities and new challenges, requiring ongoing exploration, continued refinement, and further research to address reliability, interpretability, and reproducibility.more » « less
-
Abstract Large language models (LLMs) are reshaping many aspects of materials science and chemistry research, enabling advances in molecular property prediction, materials design, scientific automation, knowledge extraction, and more. Recent developments demonstrate that the latest class of models are able to integrate structured and unstructured data, assist in hypothesis generation, and streamline research workflows. To explore the frontier of LLM capabilities across the research lifecycle, we review applications of LLMs through 32 total projects developed during the second annual LLM hackathon for applications in materials science and chemistry, a global hybrid event. These projects spanned seven key research areas: (1) molecular and material property prediction, (2) molecular and material design, (3) automation and novel interfaces, (4) scientific communication and education, (5) research data management and automation, (6) hypothesis generation and evaluation, and (7) knowledge extraction and reasoning from the scientific literature. Collectively, these applications illustrate how LLMs serve as versatile predictive models, platforms for rapid prototyping of domain-specific tools, and much more. In particular, improvements in both open source and proprietary LLM performance through the addition of reasoning, additional training data, and new techniques have expanded effectiveness, particularly in low-data environments and interdisciplinary research. As LLMs continue to improve, their integration into scientific workflows presents both new opportunities and new challenges, requiring ongoing exploration, continued refinement, and further research to address reliability, interpretability, and reproducibility.more » « less
-
Data Management Plans (DMP) are now a routine part of research proposals but are generally not referred to after funding is granted. The Belmont Forum requires an extensive document, a ‘Data and Digital Object Management Plan’ (D(DO)MP) for its awarded projects that is expected to be kept current over the life of the project. The D(DO)MP is intended to record team decisions about major tools and practices to be used over the life of the project for data and software stewardship, and for preservation of data and software products, aligned with the desired Open Science outcomes relevant to the project. Here we present one of the first instances of the use of Belmont’s D(DO)MP through a case study of the PARSEC project, a multinational and multidisciplinary investigation of the socioeconomic impacts of protected areas. We describe the development and revision of our interpretation of the D(DO)MP and discuss its adoption and acceptance by our research group. We periodically assessed the data management sophistication of team members and their use of the various nominated tools and practices. As a result, for example, we included summaries to enable the key components of the D(DO)MP to be readily viewed by the researcher. To meet the Open Science outcomes in a complex project like PARSEC, a comprehensive and appropriately structured D(DO)MP helps project leaders (a) ensure that team members are committed to the collaboration goals of the project, (b) that there is regular and effective feedback within the team, (c) training in new tools is provided as and when needed, and (d) there is easy access to a short reference to the tools and descriptions of the nominated practices.more » « less
-
SLATE (Services Layer at the Edge) is a new project that, when complete, will implement “cyberinfrastructure as code” by augmenting the canonical Science DMZ pattern with a generic, programmable, secure and trusted underlayment platform. This platform will host advanced container-centric services needed for higher-level capabilities such as data transfer nodes, software and data caches, workflow services and science gateway components. SLATE will use best-of-breed data center virtualization components, and where available, software defined networking, to enable distributed automation of deployment and service lifecycle management tasks by domain experts. As such it will simplify creation of scalable platforms that connect research teams, institutions and resources to accelerate science while reducing operational costs and development cycle times. Since SLATE will be designed to require only commodity components for its functional layers, its potential for building distributed systems should extend across all data center types and scales, thus enabling creation of ubiquitous, science-driven cyberinfrastructure. By providing automation and programmatic interfaces to distributed HPC backends and other cyberinfrastructure resources, SLATE will amplify the reach of science gateways and therefore the domain communities they support.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1039/D3DD00022B
