Research map / Federated learning and privacy / Personalized and heterogeneous FL
Federated Generative Model on Multi-Source Heterogeneous Data in IoT
Zuobin Xiong, Wei Li, Zhipeng Cai
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 |
|---|---|---|---|
| use feature-related federated generation with synchronous updating rather than the strongest reported baseline | FR data generation · MNIST and Fashion-MNIST, non-i.i.d. | FID: 14.69 -> 13.82 | Improved |
| use feature-related federated generation with synchronous updating rather than the strongest reported baseline | FR data generation · MNIST and Fashion-MNIST, non-i.i.d. | classification accuracy: 90.60% -> 90.65% | Improved |
| use asynchronous updating instead of synchronous updating (ablation) | FR data generation · MNIST and Fashion-MNIST, non-i.i.d. | FID: 13.82 -> 52.24 | Hurt |
| use asynchronous updating instead of synchronous updating (ablation) | FR data generation · MNIST and Fashion-MNIST, non-i.i.d. | classification accuracy: 90.65% -> 53.17% | Hurt |
| use label-related federated generation with synchronous updating rather than federated GAN | LR data generation · MNIST and inverse MNIST, i.i.d. | IS: 3.360 -> 3.971 | Improved |
| use label-related federated generation with synchronous updating rather than federated GAN | LR data generation · MNIST and inverse MNIST, non-i.i.d. | FID: 53.26 -> 45.22 | Improved |
| use asynchronous updating instead of synchronous updating (ablation) | LR data generation · MNIST and inverse MNIST, i.i.d. | FID: 28.56 -> 41.35 | Hurt |
| use asynchronous updating instead of synchronous updating (ablation) | LR data generation · MNIST and inverse MNIST, non-i.i.d. | IS: 3.495 -> 2.452 | Hurt |
| add domain-invariant loss to enforce shared label information across domains (ablation) | LR data generation · MNIST and inverse MNIST | same-class consistency across generated domains | Improved |
Closest papers by meaning
- Fake It Till Make It: Federated Learning with Consensus-Oriented Generation (ICLR 2024 · similarity 0.85)
- Fine-grained Control of Generative Data Augmentation in IoT Sensing (NeurIPS 2024 · similarity 0.85)
- Federated Learning with Heterogeneous LLMs: Integrating Small Student Client Models with a Large Hungry Model (AAAI 2025 · similarity 0.83)
- Towards Diverse Device Heterogeneous Federated Learning via Task Arithmetic Knowledge Integration (NeurIPS 2024 · similarity 0.83)
- An Upload-Efficient Scheme for Transferring Knowledge From a Server-Side Pre-trained Generator to Clients in Heterogeneous Federated Learning (CVPR 2024 · similarity 0.82)
- PRISM: Privacy-Preserving Improved Stochastic Masking for Federated Generative Models (ICLR 2025 · similarity 0.82)
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.
