IP23-006 VIMP: A Discourse-Aware, Community-Informed Toolkit to Predict Virality and Impact of Vaccine Misinformation Contents

IP23-006 VIMP：一个具有话语感知、社区知情的工具包，用于预测疫苗错误信息内容的病毒性和影响

• 批准号: 10762193
• 负责人: MARK Stenius ROBERTS
• 金额: $ 100万
• 依托单位: UNIVERSITY OF PITTSBURGH AT PITTSBURGH
• 依托单位国家: 美国
• 财政年份: 2024
• 资助国家: 美国
• 起止时间: 2024-02-01 至 2026-08-31
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10762193
• 关键词: IP23; 006; VIMP; Discourse; Aware

Project Title: VIMP: A Discourse-Aware, Community-Informed Toolkit to Predict Virality and Impact of Vaccine Misinformation Contents Project Abstract: We will develop a contextually grounded natural language processing (NLP) model and build a toolkit around it to leverage the befit of latest developments in data analytics, network simulations and NLP for increasing vaccine acceptance, vaccination rates and disseminating accurate scientific messages according to community needs. Our final product, VIMP, will be a discourse-aware, community-informed toolkit to predict virality and impact of vaccine misinformation contents. We will use a variety of contextually grounded NLP techniques to design our prediction models and will leverage the benefit of advanced simulation techniques for model training and evaluation. Our team will also engage with a diverse panel of community partners who provide advice on model development and tool building. CDC and its partners could use the tool to rapidly respond to misinformation about vaccines. The model and tool would further increase CDC's and its partners' ability to prioritize which misinformation content to respond to and how to allocate resources at the community level.

项目名称： VIMP：一个具有话语意识、社区知情的工具包，用于预测疫苗的病毒传播和影响 错误信息内容 项目摘要： 我们将开发一个基于上下文的自然语言处理（NLP）模型并构建一个工具包 围绕它，利用数据分析、网络模拟和自然语言处理的最新发展的优势 提高疫苗接受度、疫苗接种率并传播准确的科学信息 根据社区需求。我们的最终产品 VIMP 将是一个具有话语意识、了解社区的 预测疫苗错误信息内容的病毒性和影响的工具包。我们将使用各种 基于上下文的 NLP 技术来设计我们的预测模型，并将利用 用于模型训练和评估的先进模拟技术。我们的团队还将与 由不同的社区合作伙伴组成的小组，为模型开发和工具构建提供建议。 CDC 及其合作伙伴可以使用该工具快速应对有关疫苗的错误信息。模型和 该工具将进一步提高 CDC 及其合作伙伴优先考虑哪些错误信息内容的能力 回应以及如何在社区层面分配资源。