Research map / Federated learning and privacy / Personalized and heterogeneous FL
FedAFD: Multimodal Federated Learning via Adversarial Fusion and Distillation
Min Tan, Junchao Ma, Yinfu Feng, Jiajun Ding, Wenwen Pan, Tingting Han, Qian Zheng, Zhenzhong Kuang, Zhou Yu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add bi-level adversarial alignment (BAA) (ablation) | FedAFD · MS-COCO server (Non-IID) | rsum R@1: 59.29 -> 60.16 | Improved |
| add bi-level adversarial alignment (BAA) (ablation) | FedAFD · CIFAR-100 clients (Non-IID) | acc@1: 33.56 -> 33.18 | Hurt |
| add bi-level adversarial alignment (BAA) (ablation) | FedAFD · AGNEWS clients (Non-IID) | acc@1: 49.03 -> 51.98 | Improved |
| add granularity-aware feature fusion (GFF) (ablation) | FedAFD · CIFAR-100 clients (Non-IID) | acc@1: 24.94 -> 33.18 | Improved |
| add granularity-aware feature fusion (GFF) (ablation) | FedAFD · AGNEWS clients (Non-IID) | acc@1: 44.46 -> 51.98 | Improved |
| add granularity-aware feature fusion (GFF) (ablation) | FedAFD · MS-COCO server (Non-IID) | rsum R@1: 59.72 -> 60.16 | Improved |
| use similarity-guided ensemble distillation instead of simple averaging (ablation) | FedAFD · MS-COCO server (Non-IID) | rsum R@1: 59.56 -> 60.16 | Improved |
| use similarity-guided ensemble distillation instead of simple averaging (ablation) | FedAFD · CIFAR-100 clients (Non-IID) | acc@1: 32.21 -> 33.18 | Improved |
| increase public data from 10k to 30k (ablation) | FedAFD · MS-COCO server (Non-IID) | rsum R@1: 60.16 -> 78.09 | Improved |
| increase public data from 10k to 30k (ablation) | FedAFD · CIFAR-100 clients (Non-IID) | acc@1: 33.18 -> 30.08 | Hurt |
| increase public data from 10k to 30k (ablation) | FedAFD · Flickr30k multimodal clients i2t (Non-IID) | R@1: 32.48 -> 36.88 | Improved |
| use FedAFD instead of existing FL baselines such as CreamFL | multimodal FL with heterogeneous clients · CIFAR-100 clients (Non-IID) | acc@1: 22.14 -> 33.18 | Improved |
Closest papers by meaning
- FedMBridge: Bridgeable Multimodal Federated Learning (ICML 2024 · similarity 0.90)
- FedDAT: An Approach for Foundation Model Finetuning in Multi-Modal Heterogeneous Federated Learning (AAAI 2024 · similarity 0.90)
- Learning Reconfigurable Representations for Multimodal Federated Learning with Missing Data (NeurIPS 2025 · similarity 0.89)
- Fine-Tuning Personalization in Federated Learning to Mitigate Adversarial Clients (NeurIPS 2024 · similarity 0.89)
- Enhancing Storage and Computational Efficiency in Federated Multimodal Learning for Large-Scale Models (ICML 2024 · similarity 0.88)
- FedAS: Bridging Inconsistency in Personalized Federated Learning (CVPR 2024 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
