可边缘部署LLM用于灾难推文分类与信息价值排序
原标题：Edge-Deployable LLMs for Classification and Value-of-Information Ranking of Disaster-Related Tweets
AI 摘要
研究团队使用7个可边缘部署的小型LLM对HumAID数据集中77,484条灾难相关推文进行人道援助分类和基于ETHANE框架的信息价值（VoI）排序评估。Qwen3（8B）在10个类别中的9个取得最高F1分数，Gemma 3（4B）在VoI评估任务中多数类别召回率最高。研究表明，能在单张8GB GPU上运行的小模型可在资源受限环境下优先筛选可操作信息，用于灾难响应决策。
正文节选
- IoT - Internet of Things Edge-Deployable LLMs for Classification and Value-of-Information Ranking of Disaster-Related Tweets Thanks: Christopher Molek and K. Brent Venable were supported by NSF award NSF-RI-2008011. Abstract Humanitarian-aid and disaster relief decisions often rely on noisy, incomplete, and uncertain information from social media. Although substantial research has focused on automating the classification of such data, limited work has examined whether individual posts contain