在研课题 | DCAI-Driven Disaster Data Bank for Human Mobility Resilience - 城市之光

Research title:

A Data-Centric AI (DCAI) Driven Disaster Data Bank for Assessing Human Mobility Resilience under Extreme Events

Research Outlie:

This project focuses on the creation of a data-centric AI driven disaster data bank, investigating:

(1) The construction of a nation-wide multimodal database as a fundamental basis for digital twin applications.
(2) The development of AI-curated data features to enhance data utility and analysis.
(3) The integration of diverse data sources, including human mobility data, POI (Point of Interest) data, and disaster phenomenon data, to assess human mobility resilience.
(4) The establishment of a unified metadata framework to manage and process heterogeneous and "dirty" data with varying spatial resolutions and privacy protections.

Aim of the research:

This project aims to establish a comprehensive disaster data bank powered by Data-Centric AI (DCAI) to assess human mobility resilience during extreme events. The core objective is to integrate and curate heterogeneous data from multiple sources, such as mobile spatial statistics, POI data, and disaster phenomenon data, into a unified and high-quality database. This initiative recognizes the challenges of working with "dirty" data, which often has inconsistent spatial resolutions and is subject to privacy-preserving anonymization.

By defining unified metadata and leveraging AI for data curation, the project seeks to create a robust foundation for digital twin applications. A key consideration throughout the research will be the protection of personal confidential data, especially as data granularity increases. The ultimate goal is to generate high-quality data that can be effectively utilized to optimize disaster resilience policies and improve disaster medical operations.