New research highlights practical gap in Vertical Federated Learning backdoor defenses

A new paper on arXiv explores the practical vulnerabilities of backdoor attacks in Vertical Federated Learning (VFL). The research highlights a significant gap between theoretical findings and real-world application, noting that existing defenses often rely on unrealistic assumptions and flawed evaluation methods. To address this, the paper introduces BVBench, a new benchmark designed for practical and comprehensive evaluation of VFL backdoor risks, aiming to guide future research toward more robust defenses. AI

IMPACT Highlights practical security challenges in federated learning, potentially influencing future development of secure collaborative AI systems.

RANK_REASON The cluster contains a single academic paper published on arXiv detailing new research findings and introducing a new benchmark. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• alphaXiv
• arXiv
• BVBench
• CatalyzeX
• Connected Papers
• DagsHub
• Gotit.pub
• Hugging Face
• Litmaps
• ScienceCast
• scite Smart Citations

• paper
• safety

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →