Collaborative Research: SaTC: CORE: Medium: An Incident-Response Approach for Empowering Fact-Checkers

协作研究：SaTC：核心：媒介：增强事实检查人员能力的事件响应方法

• 批准号: 2154119
• 负责人: Chinmay Hegde
• 金额: $ 39.6万
• 依托单位: New York University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2022
• 资助国家: 美国
• 起止时间: 2022-05-01 至 2026-04-30
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2154119&HistoricalAwards=false
• 关键词: Collaborative; Research; SaTC; CORE; Medium

Fact-checking can be effective in countering the growing threat of online misinformation because people across the political spectrum and demographics tend to trust credibility judgments of fact-checkers. However, a pipeline of manual and labor-intensive practices fragmented across disparate tools makes it difficult to scale fact-checking efforts. As a result, fact-checkers are inundated with information and lack effective dissemination mechanisms for countering misinformation early and effectively. To address these challenges, this project combines the complementary information processing strengths of humans and computation to transform the efficiency, effectiveness, and scale of fact-checking. The project can enable fact-checkers to spot misinformation early, prioritize effort, and unify the various tools and techniques used for fact-checking. The research outcomes can scale the work of human fact-checkers and boost information literacy in society, which can significantly reduce the number of people exposed to misinformation.The project draws upon the core components of security incident response (i.e., preparation, detection, containment, and post-incident activity) to transform the ad-hoc, time-consuming, and small-scale nature of current fact-checking practices with a security-analyst perspective and a unified user experience (UX). The research approach leverages the power of computation and personalization while retaining the synergistic advantages of the human fact-checker in the loop. The interdisciplinary sociotechnical approach involves empirical studies of fact-checker practices, collection of data and development of computational techniques to address their challenges and barriers, and design explorations of novel UI/UX techniques to connect humans and computation. The research incorporates a feedback loop to disseminate fact-checking outcomes, thus boosting their visibility and impact on end users exposed to misinformation. The researchers are developing early warning and detection techniques to reduce the time between misinformation generation and fact-check dissemination and are employing prioritization and personalization for more effective and efficient use of fact-checking resources. The researchers are engaging with professional fact-checkers to translate the research outcomes to the real world.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

事实核查可以有效应对在线虚假信息日益增长的威胁，因为不同政治派别和人口统计的人们都倾向于相信事实核查人员的可信度判断。然而，分散在不同工具中的一系列手动和劳动密集型实践使得事实核查工作难以扩展。结果，事实核查人员被信息淹没，缺乏有效的传播机制来及早有效地反击错误信息。为了应对这些挑战，该项目结合了人类和计算的互补信息处理优势，以改变事实核查的效率、有效性和规模。该项目可以使事实核查人员尽早发现错误信息，确定工作优先级，并统一用于事实核查的各种工具和技术。研究成果可以扩大人类事实核查人员的工作范围，提高社会的信息素养，从而显着减少接触错误信息的人数。该项目利用了安全事件响应的核心组成部分（即准备、检测、遏制）和事件后活动），以安全分析师的视角和统一的用户体验 (UX) 改变当前事实核查实践的临时性、耗时性和小规模性质。该研究方法利用了计算和个性化的力量，同时保留了循环中人类事实检查者的协同优势。跨学科的社会技术方法涉及事实检查实践的实证研究、数据收集和计算技术的开发以解决其挑战和障碍，以及连接人类和计算的新颖 UI/UX 技术的设计探索。该研究采用反馈循环来传播事实核查结果，从而提高其可见性和对受到错误信息影响的最终用户的影响。研究人员正在开发早期预警和检测技术，以减少错误信息生成和事实核查传播之间的时间，并采用优先级和个性化来更有效和高效地利用事实核查资源。研究人员正在与专业的事实核查人员合作，将研究成果转化为现实世界。该奖项反映了 NSF 的法定使命，并通过使用基金会的智力价值和更广泛的影响审查标准进行评估，被认为值得支持。

项目成果

Identity-Aware Facial Age Editing Using Latent Diffusion

使用潜在扩散进行身份感知面部年龄编辑

• DOI: 10.1109/tbiom.2024.3390570
• 发表时间: 2024-01
• 期刊: and Identity Science
• 作者: Banerjee, Sudipta;Mittal, Govind;Joshi, Ameya;Mullangi, Sai Pranaswi;Hegde, Chinmay;Memon, Nasir
• 通讯作者: Memon, Nasir

True or False: Studying the Work Practices of Professional Fact-Checkers

是非：研究专业事实核查人员的工作实践

• DOI: 10.1145/3512974
• 发表时间: 2022-03
• 期刊: Proceedings of the ACM on Human-Computer Interaction
• 作者: Micallef, Nicholas;Armacost, Vivienne;Memon, Nasir;Patil, Sameer
• 通讯作者: Patil, Sameer

Fast Certification of Vision-Language Models Using Incremental Randomized Smoothing

使用增量随机平滑快速验证视觉语言模型

• 发表时间: 2024-04
• 期刊: IEEE Conference on Secure and Trustworthy Machine Learning
• 作者: Nirala, Ashutosh K;Joshi, Ameya;Sarkar, Soumik;Hegde, Chinmay
• 通讯作者: Hegde, Chinmay

Identity-Preserving Aging of Face Images via Latent Diffusion Models

通过潜在扩散模型对人脸图像进行身份保留老化

• DOI: 10.1109/ijcb57857.2023.10448860
• 发表时间: 2023-07-17
• 期刊: 2023 IEEE International Joint Conference on Biometrics (IJCB)
• 作者: Sudipta Banerjee;Govind Mittal;Ameya Joshi;C. Hegde;Nasir D. Memon
• 通讯作者: Nasir D. Memon