SaTC: CORE: Medium: Collaborative: BaitBuster 2.0: Keeping Users Away From Clickbait

SaTC：核心：媒介：协作：BaitBuster 2.0：让用户远离点击诱饵

• 批准号: 1949694
• 负责人: Matthew Wright
• 金额: $ 35.63万
• 依托单位: Rochester Institute of Tech
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-07-01 至 2025-06-30
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1949694&HistoricalAwards=false
• 关键词: SaTC; CORE; Medium; Collaborative; BaitBuster

Social media sites such as Facebook are popular platforms for spreading clickbait, links with misleading titles that do not deliver on their promises. Not only does clickbait waste users' time, it often directs users to phishing sites and sites containing spyware and malware. A large number of users fall victim to scams on social media, including those spread through clickbait, due to both a lack of awareness and a lack of appropriate warnings on social media platforms. These users are vulnerable to identity theft, online hacking, and the exposure of sensitive information to adversaries. Thus, it is critical to limit the impact of clickbait on users' security. This project is developing novel techniques to detect various forms of clickbait, especially video-based clickbait, and study user behavior on social media to design effective warning systems. The findings from this research are being incorporated into an open-source browser extension called Baitbuster 2.0, building on the original Baitbuster tool for detecting text-based clickbait. To enhance the impact of this tool, the researchers will design new training methods to raise security awareness and help users avoid clickbait in social media. The project also aims to engage underrepresented groups via outreach efforts and through developing videos to encourage women to consider cybersecurity as a career.Detecting clickbait is a major challenge, particularly as video becomes a more prominent form of media online, undermining efforts to detect misleading text. To address this challenge, the research team will take an integrated approach examining the effects of techniques used to attract clicks from users, presentation and distribution of clickbait, personalization of clickbait through crawling users’ personal information (i.e., targeted clickbait), automatic generation of face-swapping clickbait, and risk perceptions and security awareness of users. As a first step, the researchers are collecting and analyzing clickbait datasets to explore ways of identifying clickbait on social media. Using these datasets, they are developing novel applications of state-of-the-art machine learning techniques such as optical character recognition and video understanding to automatically identify video clickbait. In another thrust of this project, the researchers are studying users' clicking behavior and corresponding security mental models to better understand their vulnerability to clickbait and examine the effects of a wide range of social engineering techniques used to attract clicks from users. The findings are being used to design warning systems, which will be integrated into BaitBuster 2.0, to warn users intelligently and effectively to avoid clickbait. Finally, the usability and efficacy of the warning system and BaitBuster 2.0 are being evaluated through in-depth user studies.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.

Facebook 等社交媒体网站是传播点击诱饵的流行平台，这些链接带有不兑现承诺的误导性标题，不仅浪费用户的时间，而且经常将用户引导至包含间谍软件和恶意软件的大型网站。由于缺乏认识和社交媒体平台上缺乏适当的警告，许多用户成为社交媒体诈骗的受害者，包括通过点击诱饵传播的诈骗。这些用户很容易遭受身份盗窃、网络黑客攻击和信息泄露。敏感信息因此，限制点击诱饵对用户安全的影响至关重要，该项目正在开发新技术来检测各种形式的点击诱饵，特别是基于视频的点击诱饵，并研究社交媒体上的用户行为以设计有效的警告系统。这项研究的结果正在被纳入名为 Baitbuster 2.0 的开源浏览器扩展中，该扩展以用于检测基于文本的点击诱饵的原始 Baitbuster 工具为基础。为了增强该工具的影响，研究人员将设计新的训练方法。提高安全意识并帮助用户避免社交媒体中的点击诱饵。该项目还旨在通过外展工作和视频吸引代表性不足的群体，以鼓励女性将网络安全视为一项职业。检测点击诱饵是一项重大挑战，尤其是随着视频成为一种挑战。更突出的在线媒体形式，破坏了检测误导性文本的努力 为了应对这一挑战，研究团队将采取综合方法来检查用于吸引用户点击、点击诱饵的呈现和分发、通过爬行实现点击诱饵的个性化的技术的效果。用户的个人信息（即有针对性的点击诱饵）、自动生成换脸点击诱饵以及用户的风险认知和安全意识作为第一步，研究人员正在收集和分析点击诱饵数据集，以探索识别社交媒体上的点击诱饵的方法。利用这些数据集，他们正在开发最先进的机器学习技术的新颖应用，例如光学字符识别和视频理解，以自动识别视频标题诱饵。研究人员正在研究用户的点击行为。和相应的安全心理模型，以更好地了解其对点击诱饵的脆弱性，并检查用于吸引用户点击的各种社会工程技术的效果，研究结果将用于设计警告系统，该系统将集成到 BaitBuster 2.0 中，以发出警告。最后，通过深入的用户研究来评估预警系统和 BaitBuster 2.0 的可用性和有效性。该奖项授予 NSF 的法定使命，并通过评估认为值得支持。利用基金会的智力优势和更广泛的影响审查标准。

项目成果

A Look into User Privacy and Third-party Applications in Facebook

Facebook 中的用户隐私和第三方应用程序研究

• 发表时间: 2021
• 期刊: Information and computer security
• 影响因子: 1.4
• 作者: Seng, Sovantharith;Al-Ameen, Mahdi Nasrullah;Wright, Matthew
• 通讯作者: Wright, Matthew

Explainable Video Entailment with Grounded Visual Evidence

具有扎实视觉证据的可解释视频蕴涵

• DOI: 10.1109/iccv48922.2021.00203
• 发表时间: 2021
• 期刊: 2021 IEEE/CVF International Conference on Computer Vision (ICCV
• 作者: Chen, Junwen;Kong, Yu
• 通讯作者: Kong, Yu

A first look into users’ perceptions of facial recognition in the physical world

初步了解用户对现实世界中面部识别的看法

• DOI: 10.1016/j.cose.2021.102227
• 发表时间: 2021
• 期刊: Computers & Security
• 影响因子: 5.6
• 作者: Seng, Sovantharith;Al-Ameen, Mahdi Nasrullah;Wright, Matthew
• 通讯作者: Wright, Matthew

GateHUB: Gated History Unit with Background Suppression for Online Action Detection

• DOI: 10.1109/cvpr52688.2022.01930
• 发表时间: 2022-06
• 期刊: 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
• 作者: Junwen Chen;Gaurav Mittal;Ye Yu;Yu Kong;Mei Chen