| ID |
Name |
Period |
| DT-RICH-24-00 |
DT-RICH Site Proposal:
Digital Twin (DT) is emerging as a new paradigm of digital transformation that can virtually replicate, simulate and feed back to physical world objects for decision making in a more efficient and risk-reduced fashion. We propose to establish a unique I/UCRC for Digital Twin Research, Innovation, and Collaboration Hub (DT-RICH) to address the grand challenges of implementing the translational capability of the DT. The center’s vision is to enable effective and healthy life on Earth via digital replica of stakeholder’s concerns with the capability to predict and test potential solutions in a what-if question-answer fashion (AIST 2022). |
2024-2025 |
| DT-RICH-24-01 |
Geographic Information Retrieval by Implementing Large Language Models:
Media data, encompassing both news and social media, provides near-real-time spatiotemporal information. This data is valuable for extracting geographical locations and associated topics to gain insights into current events. Existing tools can extract topics and geographic locations separately from media content but often struggle to recognize the complex relationships between topics and geographic locations, particularly when multiple topics and numerous geographical features appear within the same paragraph. Large Language Models (LLMs), built on transformer architectures, are adept at capturing interdependencies and connections among diverse elements within data, enabling a nuanced understanding of contextual information. This project aims to leverage LLMs to classify news topics, extract geographic locations, geocode these locations, and extract date information, thereby revealing spatiotemporal patterns in near-real-time events. The techniques developed will be applied to events such as the Sudan conflicts, illegal organ trade, and the Ukraine-Russian conflicts. |
2024-2025 |
| DT-RICH-24-02 |
DT for enhanced wildfire progression, burned area detection, and air pollution analysis
This proposal endeavors to advance the comprehension, prediction, and evaluation of wildfire dynamics, burn severity, and their impact on air quality through sophisticated spatiotemporal modeling techniques. Leveraging Earth science and technical capabilities, it targets various stages of the fire lifecycle, concentrating on enhancing wildfire progression prediction, detecting burn areas and severity, and forecasting wildfire-induced air pollution. Innovative deep learning models are harnessed alongside satellite and airborne data to deepen insights into wildfire dynamics and their effects on air quality. Additionally, the proposal introduces a digital twin component to further enrich this comprehensive approach to wildfire management and mitigation. |
2024-2025 |
| DT-RICH-24-03 |
Assessing Geospatial Accessibility to Healthcare Facilities
This project aims to leverage the concept of digital twins to explore geospatial accessibility to various healthcare facilities and their impact on health outcomes. The project focuses on disparities in geospatial accessibility across the United States by taking dental care and cancer oncologists as case studies through collaboration with the American Dental Association (ADA) and Huntsman Cancer Institute (HCI). We will use different models to evaluate the spatial accessibilities in both rural and urban areas. These comparisons will also differentiate between providers that accept specific health insurance plans, such as Medicaid. The workflow that the research team will develop will also be expanded to the application of KNIME—an open-source data analytics platform—which will facilitate the replication of this analysis. Furthermore, we will create an interactive online dashboard, acting as a digital twin of the real-world scenario. This dashboard, powered by generative AI technology, will visually represent healthcare accessibility and highlight geographic disparities in accessibility. It will enable policymakers and users to engage with complex data through simple conversational queries, enhancing decision-making and promoting healthcare equity.
|
2024-2025 |
| DT-RICH-24-04 |
DT for Urban Heat Islands
This proposal addresses the problem of urban heat islands (UHI) of informal settlements in partner cities to create a common pool of strategies, solutions, and tools useful between countries and across cultures. We engage a global network of geospatial experts to engage in an innovative multi-stakeholder process where vulnerable communities identify priorities and define solutions to UHI through a data-for-development approach. One of the integral themes for the 2023 G20 is the principle of “data for development” to reimagine the built environment highlighting the use of digital technologies to address the fight against poverty and the benefits realized when digital access becomes truly inclusive and widely accessible.
Our overall framework is threefold: 1) to create a digital geospatial twin of informal settlements through mobile mapping (handheld and airborne devices) for slum upgrading to adapt to heat stress in selected partner cities; 2) to intersect state-of-the-art technologies and local knowledge to identify critical local needs for heat adaptation solutions; and 3) to establish and train local HS2 teams to build capacity for climate adaptation and resilience. The construction of digital twins of informal settlements (the physical place, the virtual counterpart, and the data connections between them) is the basis for long term data capture, maintenance, and monitoring to demonstrate their utility as part of the disaster risk reduction toolkit. This approach operationalizes existing technologies in a comprehensive and innovative way to utilize the best available data (e.g., satellite imagery and demographic data) to identify UHI. Coupled with local knowledge (e.g., access to services, locations of open space) digital twins can be assembled through mobile data collector teams to identify heat stressors (i.e., building type and materials, illegal dumping sites, impervious surfaces) and potential solutions (i.e., resilient building materials such as “cool” roofs, window types, marketplace shades, tree planting). Using the digital twin as the organizational data model, we are poised to implement an innovative mapping program to address heat solutions in informal settlements in partner cities.
|
2024-2025 |
| DT-RICH-24-05 |
Artificial Intelligence-assisted Investigation of the Association between Wildfire Smoke and Emergency Department Visits
Changes in temperature and precipitation patterns from climate change have been key drivers in increasing wildfire prevalence and severity. Beyond direct injury and death, wildfires are particularly harmful to human health since wildfire smoke is a mixture of air pollutants that can irritate the lungs, cause inflammation, and alter immune function. Investigation of the association between wildfire smoke and public health impact plays a vital role in hazard preparedness, mitigation, and public health protection. However, wildfire smoke exposure assessment remains an ongoing challenge due to the sparse and uneven distribution of in-situ monitors and the unavailability of full coverage and high spatiotemporal resolution satellite observations. This research will establish an artificial intelligence-based framework to estimate high spatiotemporal resolution ambient PM2.5 for wildfire smoke assessment and find the associations between wildfire smoke exposure and emergency department visits.
|
2024-2025 |
| DT-RICH-GMU |
Digital Twin Verification Validation and Uncertainty Quantification
Aiming to significantly improve the reliability and precision of Digital Twin (DT) models, this project proposes the development of a robust uncertainty quantification (UQ) framework. The main goal is to create a standardized workflow that incorporates uncertainty-informed methodologies into DT models to facilitate predictions with higher confidence. This is particularly important as DT models are becoming integral in sectors such as manufacturing, healthcare, and urban planning. The initiative will include a thorough comparative analysis of various UQ methods to establish best practices that enhance model performance in diverse conditions. It will also focus on accurately quantifying uncertainty levels, which is critical for enhancing risk evaluation and decision-making processes. This involves setting clearly defined confidence intervals to provide deeper insights into the predictive capabilities and limitations of DT models. Expected outcomes encompass the establishment of a methodological standard for UQ in digital twins, thereby improving strategic decisions in industries utilizing these technologies. This standardization will also provide a replicable framework for UQ integration applicable across different technological domains, enhancing the transparency and reliability of DT models in complex decision-making scenarios.
|
2024-2025 |
| DT-RICH-HVD |
Geospatial Ensemble Platform for Digital Twins
This project is to develop a foundational visual programming toolkit for geospatial analytics that supports replicable and reproducible (R&R) Digital Twins research. It aims to directly tackle the significant challenges in geospatial data analysis faced by digital twin researchers, including high learning curves with diverse software, technical barriers in interdisciplinary collaboration, and the need for standardizing data for integration and replicating analytical workflows for verification and adaptation. The project will enhance GeoAI functionalities and expand geospatial analytics capabilities within the KNIME Analytics Platform, enabling more sophisticated modeling and simulations by visual programming. Through the integration of advanced analytical features, the development of additional case studies, and the establishment of comprehensive training programs, this project aims to strengthen the toolkit for geospatial analytics, advance the application of digital twin technology in social sciences, expand the next-generation workforce, promote citizen science participation, and democratize data science to empower broader engagement in R&R Digital Twins research.
|
2024-2025 |
| DT-RICH-UW-Seattle |
Economic Modeling of the 7-1-7 Target in the United States: A Societal-Perspective Digital Twin
In history, Black Death, Smallpox, Spanish Flu and Plague of Justinian claimed round 200 million, 55 million, 40 million, and 30 million of lives, respectively; for that reason, swift and effective pandemic response capabilities are of critical importance. In this spirit, the 7-1-7 target is a strategy that may substantially contribute to saving lives. In addition, recent economic evidence suggests potential economic benefits of preventing widespread outbreaks, including economic stability from minimizing disruption to markets, preventing lockdowns, and maintaining business and investing confidence (Bloom & Zucker, 2023), (Chan, 2022) . Henceforth, incorporating information and measuring the potential economic benefits of the 7-1-7 target is pivotal for encouraging its implementation. This study seeks to quantify the potential societal value of policy implementation to achieve the 7-1-7 target. The results of our study will be presented using an interactive, online visualization tool and delivered through a scientific manuscript and a policy brief. |
2024-2025 |
| DT-RICH-UW-Madison |
A GeoAI Foundation Model for Change Detection
Foundation Models have revolutionized machine learning, leading the forefront of AI research by processing and generating content across different modalities, including images, text, audio and video. This project intends to develop a foundation model tailored for geospatial remote sensing imagery using transformer-based self-supervised learning algorithms for improving global-scale change detection, with an application focusing on crop type mapping. By incorporating multispectral remote sensing imagery, as well as spatial, environmental and socioeconomic datasets, such as the digital elevation model (DEM) and map tiles, this foundation model will encode the spectral characteristics, the spatial topology, and the (location, time) information embedded in the multimodal input data. The pretrained foundation model will then be fine-tuned with only a few labels for various downstream tasks with a focus on crop type mapping.
|
2024-2025 |
| DT-RICH-PSU |
Leveraging large multimodal model for enhanced disaster management: a case study of rapid flood mapping
Flooding is one of the most common and devastating natural hazards, leading to significant human and economic losses annually. As climate change contributes to more frequent and intense precipitation events, flooding severity is expected to increase. Rapid flood mapping provides an immediate understanding of the extent and severity of flooding. Traditional flood mapping methods such as field surveys and remote sensing, although useful, are often time-consuming and labor-intensive. Recent advancements in Artificial Intelligence (AI), especially Generative AI such as the Large Multimodal Models (LMMs) like GPT-4, offer unprecedented capabilities in analyzing vast amounts of multimodal data, providing new avenues for enhancing disaster response. This proposal aims to develop a novel approach using Generative AI to enhance disaster management through rapid flood mapping. Our goal is to efficiently collect, process and analyze flood-related multimodal data (images, videos, and texts) from various sources such as social media and news reports to create accurate and real-time flood inundation maps, which are crucial for disaster response. Specifically, we will use generative AI to 1) gather flood-related multimodal data from various platforms including Twitter (X) and TikTok, 2) estimate flood water depth from the collected photos/videos, 3) extract the geo-location information at the street level for the flooded area shown on the photos/videos, and 4) develop a flood mapping model by integrating the extracted flood water depth, flood location, DEM, and other available datasets (such as satellite images) to produce the inundation maps.
|
2024-2025 |
| DT-RICH-Emory |
Artificial Intelligence for Digital Twins in spatiotemporal prediction and knowledge discovery
Remote sensing data has become an indispensable means to study relevant problems like space weather prediction, environmental monitoring, and disaster management, thanks to the vantage points from space. Over the years, extensive scientific research has amassed a wealth of knowledge in these fields, yet this knowledge remains insufficient to handle the complex mechanisms of these events and the rapidly increasing volume of observations. Recent advancements in Artificial Intelligence (AI), including machine learning and advanced image processing, have been applied to improve the predictability of these events by harnessing vast amounts of accumulated data. However, the current AI methods often lack transparency and cannot align with verified scientific knowledge. To fill this gap, this project aims at a generic framework that synergizes the strength of AI methods and human scientific knowledge by proposing transformative techniques that discover knowledge from AI models, guide AI models with existing knowledge, and an interactive virtuous circle between AI and scientists. This project will build an integrated end-to-end workflow of AI for remote sensing data, tackle knowledge discovery, and alleviate the communication barrier between domain experts and AI models.
|
2024-2025 |
| DT-RICH-VT |
Physician Digital Twins for Equitable Workload Management, Improved Quality of Care, and Burnout Prevention
Healthcare systems (HSs) are complex socio-technical systems that rely on seamless collaboration between healthcare providers (i.e., doctors and nurses) with distinct expertise and engineered systems (e.g., electronic health records) to provide safety-critical services for communities. While provider performance plays a crucial role in the overall functionality and effectiveness of the broader HS; provider wellbeing has been consistently overlooked in the U.S. and this trend led to systemic issues such as burnout as documented by the National Academies. Burnout is detrimental to both the HSs and the providers suffering from it. On the HS-level, it leads to decreased quality of care and patient satisfaction, along with increased safety incidents, patient mortality, and operational costs. On a provider-level, it leads to a deterioration in mental health, increased likelihood of substance abuse and occupational injuries; thus, perpetuating a vicious cycle that negatively impact provider wellbeing. To that end, this research leverages the advancing digital twin (DT) technologies, particularly their real-time monitoring and intervention capabilities, to address the provider burnout issue at its core by developing a prototype DT of an existing microsystem in healthcare, a family medicine clinic operated by Carilion. This project will utilize systems engineering techniques to model the provider-technology-patient interplay and then use data-driven methods to quantify provider workload. Formulated workload models will be empirically verified through an array of mixed-methods approaches, and the resulting DT will be implemented into practice. By doing so, this project concurrently addresses the knowledge gaps (i) in the DT literature regarding representation of human decision-makers in the loop, (ii) the virtual-to-physical mapping challenges that have been documented in DTs for HSs, (iii) agile measurement and mitigation of burnout in HSs research through rigorous analytical techniques that are transferable to other representative HSs.
|
2024-2025 |
| DT-RICH-HVD-L1 |
Standards and Standardization Processes as an International Channel for Global Knowledge Transfer
This project investigates the dissemination of scientific knowledge through international standards within companies in Europe, the US, and China. It aims to identify the role of standards in the concrete application of research-generated knowledge, highlighting the recent recognition of their significance in knowledge and technology transfer (KTT). Building on a previous study focused on Germany, this project addresses the international dimension of such dissemination networks. Through analyzing over 25 million corporate websites, the project seeks to explore the adoption and diffusion of standards as indicators of global KTT. Key questions include the incorporation of global scientific knowledge into standards, the factors influencing KTT via standards, and the macroeconomic implications for global trade and technology sovereignty. This initiative contributes to understanding the strategic role of standards in global knowledge flows and innovation, emphasizing the need for further integration of standardization in KTT metrics and policies. |
2024-2025 |
| DT-RICH-HVD-L2 |
Developing a Tactical Thermal Transition Tool to Support the Efficient and Equitable Conversion of Gas to Geothermal Network Infrastructure
Buildings account for 35% of greenhouse gas emissions in Massachusetts, making them the second largest source of emissions in the state. Most of these emissions come from space and water heating systems. Gas is the most common heating source in the state, with 52% of households using natural gas, followed by 22% using fuel oil and kerosene. Electricity is used in 18% of the homes. Geothermal networks provide an efficient way to electrify heating at a societal scale and reduce greenhouse gas emissions from the built environment sector. A recent study by Oak Ridge National Laboratory found that if ground-source heat pumps were installed in 70% of US buildings, they would reduce the electricity generation capacity by 410 GW and eliminate the need for installing 43,500 miles of transmission lines. This project will develop a proof of concept for a gas-to geo network digital twin which can be used to optimize the transition from natural gas to geothermal networks within the City of Cambridge, Massachusetts. The goal of the twin is to enable the City to reduce costs, time, and disruption of the transition while prioritizing equity. If this work is successful it could be useful in the transition of other cities and potentially states. |
2024-2025 |
| DT-RICH-HVD-L3 |
Spatial Structures of Democracy-Threatening Online Communication
Highly polarizing online users and emotionalized content pose a significant threat to democracies. While information diffusion in social networks has been studied extensively, the geospatial diffusion of democracy-threatening online communication (DTOC) remains under researched. The proposal at hand aims to address this research gap and to gain insight into the spatial diffusion process using geo-referenced social media data. The project proposes a multimodal machine learning (ML) methodology that considers the network, semantic, and emotion modalities of geo-social media data (geospatial AI – geoAI). Together with the investigation of explainable AI (XAI) techniques, this improves the understanding of drivers in the spatial diffusion of DTOC. The project follows a mixed-method approach (exploratory and explanatory) at the intersection of geoinformatics and humanities (communication and political sciences), generating insights about the origins, drivers, and the spatial diffusion of democracy-threatening online communication.
|
2024-2025 |
