Research map / Federated learning and privacy / Personalized and heterogeneous FL
Multi-Modal Style Transfer-based Prompt Tuning for Efficient Federated Domain Generalization
Yuliang Chen, Xi Lin, Jun Wu, Xiangrui Cai, Qiaolun Zhang, Xichun Fan, Jiapeng Xu, Xiu Su
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 multi-modal style transfer (MST) to global-prompt variant (ablation) | FaST-PT · PACS ablation (average over 4 target domains) | average accuracy: 92.58 -> 94.38 | Improved |
| add MST to global+domain prompts with DPG (ablation) | FaST-PT · PACS ablation (average over 4 target domains) | average accuracy: 97.06 -> 98.16 | Improved |
| add contrastive loss between global and domain prompts (ablation) | FaST-PT · PACS ablation (average over 4 target domains) | average accuracy: 97.95 -> 98.16 | Improved |
| combine global and domain prompts instead of either alone (ablation) | FaST-PT · PACS ablation (average over 4 target domains) | average accuracy: 87.67 -> 97.06 | Improved |
| add target-domain text description for augmentation (ablation) | FaST-PT · PACS ablation (average over 4 target domains) | average accuracy: 98.16 -> 98.42 | No clear effect |
| increase prompt length beyond 4 (ablation) | FaST-PT · PACS and DomainNet | accuracy | Mixed |
| FaST-PT vs DiPrompt | FaST-PT · PACS average | accuracy: 94.86 -> 98.16 | Improved |
| FaST-PT vs DiPrompt | FaST-PT · DomainNet average | accuracy: 86.84 -> 89.17 | Improved |
| FaST-PT vs DiPrompt | FaST-PT · VLCS C domain | accuracy: 99.70 -> 99.17 | Hurt |
| use prompt exchange instead of full model upload | FaST-PT · communication | uploaded parameters: 305M -> 0.005M | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DiPrompt — DiPrompT: Disentangled Prompt Tuning for Multiple Latent Domain Generalization in Federated Learning (CVPR 2024) · better on 6 of 6 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- DiPrompT: Disentangled Prompt Tuning for Multiple Latent Domain Generalization in Federated Learning (CVPR 2024 · similarity 0.91)
- Federated Adaptive Prompt Tuning for Multi-Domain Collaborative Learning (AAAI 2024 · similarity 0.90)
- Global Prompt Refinement with Non-Interfering Attention Masking for One-Shot Federated Learning (NeurIPS 2025 · similarity 0.88)
- TTA-FedDG: Leveraging Test-Time Adaptation to Address Federated Domain Generalization (AAAI 2025 · similarity 0.88)
- Harmonizing Generalization and Personalization in Federated Prompt Learning (ICML 2024 · similarity 0.88)
- FedPHA: Federated Prompt Learning for Heterogeneous Client Adaptation (ICML 2025 · 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.
