CSAMGuard: Leveraging Advanced Machine Learning to Protect Against CSAM Link Obfuscation

CSAMGuard：利用先进的机器学习来防止 CSAM 链接混淆

• 批准号: 10073540
• 金额: $ 15.29万
• 依托单位: CENTRE FOR FACTORIES OF THE FUTURE LIMITED
• 依托单位国家: 英国
• 项目类别: Collaborative R&D
• 财政年份: 2023
• 资助国家: 英国
• 起止时间: 2023 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=10073540
• 关键词: CSAMGuard; Leveraging; Advanced; Machine; Learning

The CSAMGuard project presents an innovative approach to combat Child Sexual Abuse Material (CSAM) by focusing on the detection and prevention of CSAM link shortening and modification through the use of a machine learning (ML) model. This method promises enhanced accuracy and a reduction in false positives compared to traditional approaches like blacklisting or heuristic methods.Our central focus is on the development and implementation of the Central CSAM Intelligence System (CCIS), a sophisticated and specialized system dedicated to identifying and disrupting CSAM link shortening and modification. The CCIS incorporates multiple key services to address the CSAM challenge, including a CSAM Scanner designed to identify potential CSAM links, a CSAM Blocker to prevent access to them, and a CSAM Reporter responsible for notifying relevant authorities, including Link Shortening Service Providers.

CSAMGuard项目提出了一种打击儿童性虐待材料（CSAM）的创新方法，重点是通过使用机器学习（ML）模型来检测和预防CSAM链接缩短和修改。与传统的黑名单或启发式方法相比，这种方法可以提高准确性并减少误报。我们的中心重点是中央CSAM智能系统（CCIS）的开发和实施，这是一个复杂而专业的系统，致力于识别和破坏CSAM链路缩短和修改。CCIS整合了多项关键服务来应对CSAM挑战，包括旨在识别潜在CSAM链接的CSAM扫描器，阻止访问它们的CSAM阻止器，以及负责通知相关当局的CSAM报告器，包括链路缩短服务提供商。