Enhancing emergency decision-making with knowledge graphs and large language models

International Journal of Disaster Risk Reduction, Год журнала: 2024, Номер unknown, С. 104804 - 104804

Опубликована: Сен. 1, 2024

Язык: Английский

---

Text mining of practical disaster reports: Case study on Cascadia earthquake preparedness

PLoS ONE, Год журнала: 2025, Номер 20(1), С. e0313259 - e0313259

Опубликована: Янв. 7, 2025

Many practical disaster reports are published daily worldwide in various forms, including after-action reports, response plans, impact assessments, and resiliency plans. These serve as vital resources, allowing future generations to learn from past events better mitigate prepare for disasters. However, this extensive literature often has limited on research practice due challenges synthesizing analyzing the reports. In study, we 1) present a corpus of text mining 2) introduce an approach extract insights using select tools. We validate through case study examining preparedness U.S. Pacific Northwest magnitude 9 Cascadia Subduction Zone earthquake, which potential disrupt lifeline infrastructures months. To explore opportunities associated with conducted brief survey user groups. The illustrates types that our can corpus. Notably, it reveals differences priorities between Washington Oregon state-level emergency management, uncovers latent sentiments expressed within identifies inconsistent vocabulary across field. Survey results highlight while simple tools may yield primarily interpretable by experienced professionals, more advanced utilizing large language models, such Generative Pre-trained Transformer (GPT), offer accessible insights, albeit known risk current artificial intelligence technologies. ensure reproducibility, all supporting data code made publicly available (DOI: 10.17603/ds2-9s7w-9694 ).

Язык: Английский

---

Can Large Language Models Effectively Reason about Adverse Weather Conditions?

Environmental Modelling & Software, Год журнала: 2025, Номер unknown, С. 106421 - 106421

Опубликована: Март 1, 2025

Язык: Английский

---

A Multidimensional Study of the 2023 Beijing Extreme Rainfall: Theme, Location, and Sentiment Based on Social Media Data

ISPRS International Journal of Geo-Information, Год журнала: 2025, Номер 14(4), С. 136 - 136

Опубликована: Март 24, 2025

Extreme rainfall events are significant manifestations of climate change, causing substantial impacts on urban infrastructure and public life. This study takes the extreme event in Beijing 2023 as background utilizes data from Sina Weibo. Based large language models prompt engineering, disaster information is extracted, a multi-factor coupled multi-sentiment classification model, Bert-BiLSTM, designed. A analysis framework focusing three dimensions theme, location sentiment constructed. The results indicate that during pre-disaster stage, themes concentrated warnings prevention, shifting to specific rescue actions disaster, post-disaster, they express gratitude personnel highlight social cohesion. In terms spatial location, shows clustering, predominantly occurring Mentougou Fangshan. There clear difference emotional expression between official media public; primarily focuses neutral reporting fact dissemination, while even richer. At same time, there also variations expressions across different affected regions. provides new perspectives methods for analyzing by revealing evolution themes, distribution disasters, temporal changes sentiment. These insights can support risk assessment, resource allocation, opinion guidance emergency management, thereby enhancing precision effectiveness response strategies.

Язык: Английский

---

A Cross-Border Biorisk Toolkit for Healthcare Professionals

International Journal of Environmental Research and Public Health, Год журнала: 2024, Номер 21(9), С. 1261 - 1261

Опубликована: Сен. 23, 2024

The COVID-19 pandemic posed significant challenges to public health, exposing first responders high biosafety risks during medical assistance and containment efforts. PANDEM-2 study aimed address these critical issues by emphasising the importance of frequently updated, harmonised guidelines. This reviewed scientific publications, lessons learned, real-world experiences from identify biorisk gaps in three areas: (i) patient transportation management, (ii) sample handling testing, (iii) data management communication laboratory staff. At onset pandemic, faced several challenges, including rapid expansion emergency services, conversion non-medical structures, increased internal cross-border transport infected patients, frequent changes protocols, a shortage personal protective equipment. In response, this developed versatile easily adaptable toolkit, guidance recommendations linked updated national international online repositories. It establishes groundwork for minimum standard that can be tailored various response scenarios, using monkeypox as fictive test case. toolkit enables access information via QR codes mobile devices, improving providing an standardised approach caregivers involved responses.

Язык: Английский

---

Ethics of Using LLMs in Content Moderation on Twitter

International Journal of Innovative Science and Research Technology (IJISRT), Год журнала: 2024, Номер unknown, С. 2487 - 2493

Опубликована: Ноя. 16, 2024

As the number of users increases on social media each year, posts that are made rises gradually. This is relevant for with negative characters including hate speech, misinformation, explicit material, or cyberbullying influences terribly users’ experience. paper puts emphasis content moderation LLMs to avoid issues bias, transparency, free and accountability. Several experiments were conducted pre-trained models identify efficiency arising ethical concerns while moderating posted data. Our findings reveal demonstrate bias during from different demographics minority communities. One most significant challenges found was lack transparency in LLM's decision-making process. Despite concerns, LLM demonstrated processing large volumes content, this significantly reduced time required flag potentially harmful posts. research highlights need a balanced approach protecting freedom speech ensuring responsible use NLP online platforms.

Язык: Английский

---

Gen-AI for User Safety: A Survey

2021 IEEE International Conference on Big Data (Big Data), Год журнала: 2024, Номер unknown, С. 5315 - 5324

Опубликована: Дек. 15, 2024

Язык: Английский

---

Short Paper: AI-Driven Disaster Warning System: Integrating Predictive Data with LLM for Contextualized Guideline Generation

Опубликована: Дек. 19, 2024

Язык: Английский

---