Scholarlink

The adoption of federated learning (FL) in digital forensics enables collaborative analysis while preserving data privacy. However, its decentralized nature introduces significant cybersecurity risks. Cyber threat intelligence (CTI) plays a crucial role in identifying, assessing, and mitigating these threats, helping forensic investigators predict vulnerabilities and design effective defenses for FL systems. This chapter explores the intersection of CTI and FL in digital forensics, outlining key FL vulnerabilities in communication channels, participant models, and data. It categorizes attacks-targeted, untargeted, and backdoor-based on mode, status, and setting, providing a structured view of potential threats. Additionally, the chapter examines security risks unique to different FL types: feature inference in horizontal FL (HFL), label inference in vertical FL (VFL), and utility threats in federated transfer learning (FTL). By analyzing these threats and emphasizing CTI-driven security strategies, the chapter highlights the necessity of robust protections to ensure the integrity and confidentiality of forensic FL systems.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Cyber Threat Intelligence and Security for Federated Learning in Digital Forensics

  • 章节

  • 发表日期：

    2025年

查看全文
  • S. S. Iyengar,
  • Seyedsina Nabavirazavi,
  • Yashas Hariprasad,
  • Prasad HB,
  • C. Krishna Mohan

摘要

The adoption of federated learning (FL) in digital forensics enables collaborative analysis while preserving data privacy. However, its decentralized nature introduces significant cybersecurity risks. Cyber threat intelligence (CTI) plays a crucial role in identifying, assessing, and mitigating these threats, helping forensic investigators predict vulnerabilities and design effective defenses for FL systems. This chapter explores the intersection of CTI and FL in digital forensics, outlining key FL vulnerabilities in communication channels, participant models, and data. It categorizes attacks-targeted, untargeted, and backdoor-based on mode, status, and setting, providing a structured view of potential threats. Additionally, the chapter examines security risks unique to different FL types: feature inference in horizontal FL (HFL), label inference in vertical FL (VFL), and utility threats in federated transfer learning (FTL). By analyzing these threats and emphasizing CTI-driven security strategies, the chapter highlights the necessity of robust protections to ensure the integrity and confidentiality of forensic FL systems.