[2512.18809] FedVideoMAE: Efficient Privacy-Preserving Federated Video Moderation

Computer Science > Computer Vision and Pattern Recognition arXiv:2512.18809 (cs) [Submitted on 21 Dec 2025 (v1), last revised 3 Apr 2026 (this version, v2)] Title:FedVideoMAE: Efficient Privacy-Preserving Federated Video Moderation Authors:Ziyuan Tao, Chuanzhi Xu, Sandaru Jayawardana, Adnan Mahmood, Wei Bao, Kanchana Thilakarathna, Teng Joon Lim View a PDF of the paper titled FedVideoMAE: Efficient Privacy-Preserving Federated Video Moderation, by Ziyuan Tao and 6 other authors View PDF HTML (experimental) Abstract:Short-form video moderation increasingly needs learning pipelines that protect user privacy without paying the full bandwidth and latency cost of cloud-centralized inference. We present FedVideoMAE, an on-device federated framework for video violence detection that combines self-supervised VideoMAE representations, LoRA-based parameter-efficient adaptation, client-side DP-SGD, and server-side secure aggregation. By updating only 5.5M parameters (about 3.5% of a 156M backbone), FedVideoMAE reduces communication by 28.3x relative to full-model federated updates while keeping raw videos on device throughout training. On RWF-2000 with 40 clients, the method reaches 77.25% accuracy without privacy protection and 65~66% under strong differential privacy. We further show that this privacy gap is consistent with an effective-SNR analysis tailored to the small-data, parameter-efficient federated regime, which indicates roughly 8.5~12x DP-noise amplification in our settin...