SCC-PG: JST: Multimodal Data Analytics and Integration for Emergency Response and Disaster Management

SCC-PG：JST：应急响应和灾害管理的多模式数据分析和集成

• 批准号: 1952089
• 负责人: Shu-Ching Chen
• 金额: $ 7.5万
• 依托单位: Florida International University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-07-01 至 2022-06-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1952089&HistoricalAwards=false
• 关键词: SCC; PG; JST; Multimodal; Data

The prevalence of high-speed data management and communication systems have produced countless large volumes of high-veracity, real-time data collections across many domains. For example, natural disasters initiate a myriad of human communications, data exchange, and situational assessments. It is crucial to quickly collect and analyze all of the relevant data, as life or death decisions may rest on the outcome. Hurricane Irma in 2017 caused the largest civil evacuation in Florida’s history and emphasizes two critical problems in emergency management: 1) pre-storm evacuation coordination; and 2) post-storm support for shelter-in-place locations. Disaster information integration and fusion technologies have the potential to deliver enhanced situational awareness tools across all sectors, enabling a more efficient, effective, and automated emergency management and recovery process. However, there is no discernible approach to identify and exploit the underlying patterns of each dataset while also minimizing possible drawbacks. This project aims to utilize multimodal data, such as text messages, images, videos, traffic information, and geo-referencing information, from various sources including social media, news, government announcements, and radio broadcasts. This work will investigate the analysis and fusion of this information to provide useful insights for aiding the decision-making process for both residents and government agencies. The goal is to develop new tools and technologies that can support emergency managers to better evaluate the effectiveness of disaster management policies such as evacuation. The proposed research provides potential solutions to solve crucial information analysis challenges related to disaster information management while leveraging the team's previous work. In addition, the team's research approach offers rapid key information identification, efficient multimodal data integration that facilitates emergency management, and enhances dynamic community disaster information sharing. Moreover, solutions developed could later be extended to other domains in the information management field. This project fosters collaboration among two institutions, Florida International University (FIU) and University of Tokyo, as well as institutions across the public and private sectors, to develop advanced techniques for effective emergency response and disaster management. The broader impact of this work will lead to scientific advances in the preparation, response, recovery, and mitigation of major disasters. As the largest graduate Hispanic Serving Institution in the continental United States, FIU will also benefit from the impact of this project that expands the participation of underrepresented groups in STEM fields. The research findings of this project will be broadly disseminated via publications, presentations, and an organized workshop.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

高速数据管理和通信系统的流行已经在许多领域产生了无数的高准确性、实时的数据收集。例如，自然灾害引发了无数的人类交流、数据交换和情况评估。快速收集和分析所有相关数据至关重要，因为生死攸关的决定可能取决于结果。2017年的飓风伊尔玛造成了佛罗里达州历史上最大规模的民用疏散，并强调了应急管理中的两个关键问题：1)风暴前的疏散协调；2)风暴后对原地避难所的支持。灾害信息整合和融合技术有可能在所有部门提供增强的态势感知工具，从而实现更高效、更有效和更自动化的应急管理和恢复过程。然而，没有可识别的方法来识别和利用每个数据集的基本模式，同时还将可能的缺点降至最低。该项目旨在利用来自各种来源的多模式数据，如文本消息、图像、视频、交通信息和地理参考信息，包括社交媒体、新闻、政府公告和无线电广播。这项工作将调查这些信息的分析和融合，以提供有用的见解，以帮助居民和政府机构的决策过程。目标是开发新的工具和技术，以支持应急管理人员更好地评估疏散等灾害管理政策的有效性。拟议的研究提供了潜在的解决方案，以解决与灾害信息管理有关的关键信息分析挑战，同时利用该小组以前的工作。此外，该团队的研究方法提供了快速的关键信息识别、高效的多模式数据集成，以促进应急管理，并增强动态的社区灾害信息共享。此外，制定的解决方案以后可以扩展到信息管理领域的其他领域。该项目促进了佛罗里达国际大学(FIU)和东京大学两个机构以及公共和私营部门的机构之间的合作，以开发有效的应急和灾害管理的先进技术。这项工作的更广泛影响将导致重大灾害的准备、应对、恢复和减轻方面的科学进步。作为美国大陆最大的拉美裔毕业生服务机构，FIU也将受益于这一项目的影响，该项目扩大了STEM领域代表性不足群体的参与。该项目的研究成果将通过出版物、演示文稿和有组织的工作坊广泛传播。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

项目成果

Multi-Label Multi-Task Learning with Dynamic Task Weight Balancing

• DOI: 10.1109/iri49571.2020.00042
• 发表时间: 2020-08
• 期刊: 2020 IEEE 21st International Conference on Information Reuse and Integration for Data Science (IRI)
• 作者: Tianyi Wang;Shu‐Ching Chen

Data Analytics for Air Travel Data: A Survey and New Perspectives

• DOI: 10.1145/3469028
• 发表时间: 2021-11-01
• 期刊: ACM COMPUTING SURVEYS
• 影响因子: 16.6
• 作者: Tian, Haiman;Presa-Reyes, Maria;Iyengar, Sundaraja Sitharama
• 通讯作者: Iyengar, Sundaraja Sitharama

Weakly-Supervised Damaged Building Localization and Assessment with Noise Regularization

弱监督受损建筑定位和噪声正则化评估

• DOI: 10.1109/mipr51284.2021.00009
• 发表时间: 2021
• 期刊: The 4th IEEE International Conference on Multimedia Information Processing and Retrieval
• 作者: Presa-Reyes, Maria;Chen, Shu-Ching
• 通讯作者: Chen, Shu-Ching

Hierarchical Multimodal Fusion Network with Dynamic Multi-task Learning

• DOI: 10.1109/iri51335.2021.00034
• 发表时间: 2021-08
• 期刊: 2021 IEEE 22nd International Conference on Information Reuse and Integration for Data Science (IRI)
• 作者: Tianyi Wang;Shu‐Ching Chen

Data-Driven In-Crisis Community Identification for Disaster Response and Management

• DOI: 10.1109/cic52973.2021.00021
• 发表时间: 2021-12
• 期刊: 2021 IEEE 7th International Conference on Collaboration and Internet Computing (CIC)
• 作者: Yudong Tao;Renhe Jiang;Erik Coltey;Chuang Yang;Xuan Song;R. Shibasaki;Mei-Ling Shyu;Shu‐Ching Chen