FedVideoMAE: Efficient Privacy-Preserving Federated Video Moderation
This addresses privacy risks and efficiency issues in video moderation for short-form platforms, though it is incremental as it builds on existing federated learning and self-supervised methods.
The paper tackles privacy-preserving video moderation by proposing an on-device federated learning framework for video violence detection, achieving 77.25% accuracy without privacy and 65-66% under strong differential privacy while reducing communication cost by 28.3x.
The rapid growth of short-form video platforms increases the need for privacy-preserving moderation, as cloud-based pipelines expose raw videos to privacy risks, high bandwidth costs, and inference latency. To address these challenges, we propose an on-device federated learning framework for video violence detection that integrates self-supervised VideoMAE representations, LoRA-based parameter-efficient adaptation, and defense-in-depth privacy protection. Our approach reduces the trainable parameter count to 5.5M (~3.5% of a 156M backbone) and incorporates DP-SGD with configurable privacy budgets and secure aggregation. Experiments on RWF-2000 with 40 clients achieve 77.25% accuracy without privacy protection and 65-66% under strong differential privacy, while reducing communication cost by $28.3\times$ compared to full-model federated learning. The code is available at: {https://github.com/zyt-599/FedVideoMAE}