Research map / Federated learning and privacy / Personalized and heterogeneous FL
Toward Enhancing Representation Learning in Federated Multi-Task Settings
Mehdi Setayesh, Mahdi Beitollahi, Yasser H. Khalil, Hongliang Li
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 |
|---|---|---|---|
| align all models jointly with Muscle loss rather than baselines | FedMuscle · Setup1 Pascal VOC public | Delta (%): +5.83 (CoFED) -> +26.70 | Improved |
| use Muscle loss instead of Gramian-based contrastive loss (ablation) | FedMuscle · Setup1 Pascal VOC, COCO, CIFAR-100 | Delta (%): +11.2%, +28.4%, +11.1% | Improved |
| use pairwise alignment instead of Muscle loss within FedMuscle (ablation) | FedMuscle · Setup1 | Delta (%) | Hurt |
| replace CreamFL contrastive regularization with Muscle loss | CreamFL · 1K and 5K test images retrieval | Delta (%): +0.93 -> +1.17 (1K); +0.88 -> +1.94 (5K) | Improved |
| use public dataset with detailed images (Pascal VOC, COCO) instead of CIFAR-100 (ablation) | FedMuscle · Setup1 | Delta (%): +16.88 -> +26.70 / +28.65 | Improved |
| increase number of selected representation matrices M from 3 to 5 (ablation) | FedMuscle · Setup1 Pascal VOC | Delta (%): +26.70 -> +27.74 | No clear effect |
| increase M from 1 to 3 (ablation) | FedMuscle · Setup1 Pascal VOC | Delta (%): +17.90 -> +26.70 | Improved |
| add text and segmentation users (Setup2) | FedMuscle · Setup2 DistilBERT TC user | accuracy: 56.03 -> 54.10 | Hurt |
| increase number of communication rounds (ablation) | FedMuscle · Setup1 COCO public | Delta (%) | Improved |
| increase CL epochs at low communication rounds (ablation) | FedMuscle · Setup1 COCO public | Delta (%) | Improved |
| increase batch size (ablation) | FedMuscle · Setup1 COCO public | Delta (%) | Improved |
Closest papers by meaning
- A Mutual Information Perspective on Federated Contrastive Learning (ICLR 2024 · similarity 0.87)
- FedHCA2: Towards Hetero-Client Federated Multi-Task Learning (CVPR 2024 · similarity 0.87)
- Federated Model Heterogeneous Matryoshka Representation Learning (NeurIPS 2024 · similarity 0.87)
- Global and Local Prompts Cooperation via Optimal Transport for Federated Learning (CVPR 2024 · similarity 0.86)
- Learning Task-Agnostic Representations through Multi-Teacher Distillation (NeurIPS 2025 · similarity 0.85)
- Enhancing Storage and Computational Efficiency in Federated Multimodal Learning for Large-Scale Models (ICML 2024 · similarity 0.85)
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.
