LLMs as World Models: Data-Driven and Human-Centered Pre-Event Simulation for Disaster Impact Assessment
University of South Florida（佛罗里达州立大学）；Arizona State University（亚利桑那州立大学）；Massachusetts Institute of Technology（麻省理工学院）；New York University（纽约大学）；University of Alabama（阿拉巴马大学）；Texas Tech University（德克萨斯科技大学）
Lingyao Li, Dawei Li, Zhenhui Ou, Xiaoran Xu, Jingxiao Liu, Zihui Ma, Runlong Yu, Min Deng
Efficient simulation is essential for enhancing proactive preparedness for sudden-onset disasters such as earthquakes. Recent advancements in large language models (LLMs) as world models show promise in simulating complex scenarios. This study examines multiple LLMs to proactively estimate perceived earthquake impacts. Leveraging multimodal datasets including geospatial, socioeconomic, building, and street-level imagery data, our framework generates Modified Mercalli Intensity (MMI) predictions at zip code and county scales. Evaluations on the 2014 Napa and 2019 Ridgecrest earthquakes using USGS ''Did You Feel It? (DYFI)'' reports demonstrate significant alignment, as evidenced by a high correlation of 0.88 and a low RMSE of 0.77 as compared to real reports at the zip code level. Techniques such as RAG and ICL can improve simulation performance, while visual inputs notably enhance accuracy compared to structured numerical data alone. These findings show the promise of LLMs in simulating disaster impacts that can help strengthen pre-event planning.
A Multimodal, Multilingual, and Multidimensional Pipeline for Fine-grained Crowdsourcing Earthquake Damage Evaluation
New York University（纽约大学）；University of South Florida（佛罗里达州立大学）；Google LLC（谷歌公司）；University of California, Santa Barbara（加州大学圣巴巴拉分校）；Massachusetts Institute of Technology（麻省理工学院）
Toward a Human-Centered Evaluation Framework for Trustworthy LLM-Powered GUI Agents
University of Notre Dame（圣约翰大学）；Northeastern University（东北大学）；Virginia Tech（弗吉尼亚理工大学）；University of Washington（华盛顿大学）；Johns Hopkins University（约翰霍普金斯大学）
Assessing care quality and performance is essential to improve healthcare processes and population health management. However, due to bad system design and lack of access to required data, this assessment is often delayed or not done at all. The goal of our research is to investigate an advanced analytics platform that enables healthcare quality and performance assessment. We used a user-centered design approach to identify the system requirements and have the concept of episode of care as the building block of information for a key performance indicator analytics system. We implemented archit…
The objective of this study is to examine spatial patterns of impacts and recovery of communities based on variances in credit card transactions. Such variances could capture the collective effects of household impacts, disrupted accesses, and business closures, and thus provide an integrative measure for examining disaster impacts and community recovery in disasters. Existing studies depend mainly on survey and sociodemographic data for disaster impacts and recovery effort evaluations, although such data has limitations, including large data collection efforts and delayed timeliness results. …
Developing a Multi-Agent System to Generate Next Generation Science Assessments with Evidence-Centered Design
AI4STEM Education Center, University of Georgia（AI4STEM教育中心，佐治亚大学）；Department of Educational Psychology, University of Georgia（教育心理学系，佐治亚大学）；School of Computing, University of Georgia（计算学院，佐治亚大学）；Department of Mathematics, Science, and Social Studies Education, University of Georgia（数学、科学与社会科学教育系，佐治亚大学）
AB2CD: AI for Building Climate Damage Classification and Detection
IBM Consulting, Germany（IBM咨询公司（德国））；Karlsruhe Institute of Technology, Germany（卡尔斯鲁厄理工学院）；IBM Research – Europe, Switzerland（IBM欧洲研究院）；University of Bayreuth, Germany（拜罗伊特大学）
Facebook Disaster Maps (FBDM) is the first platform providing analysis-ready population change products derived from crowdsourced data targeting disaster relief practices. We evaluate the representativeness of FBDM data using the Mann-Kendall test and emerging hot and cold spots in an anomaly analysis to reveal the trend, magnitude, and agglommeration of population displacement during the Mendocino Complex and Woolsey fires in California, USA. Our results show that the distribution of FBDM pre-crisis users fits well with the total population from different sources. Due to usage habits, the eld…
