Distilling A Universal Expert from Clustered Federated Learning
Authors: Zeqi Leng, Chunxu Zhang, Guodong Long, Riting Xia, Bo Yang
IJCAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experimental results demonstrate the superior performance of the proposed method across various scenarios, highlighting its potential to advance the state of CFL by balancing personalized and shared knowledge more effectively. 5 Experiments |
| Researcher Affiliation | Academia | 1Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education, China 2College of Computer Science and Technology, Jilin University, China 3Australian Artificial Intelligence Institute, FEIT, University of Technology Sydney 4College of Computer Science, Inner Mongolia University, Hohhot, China EMAIL, EMAIL, EMAIL EMAIL, EMAIL |
| Pseudocode | Yes | Algorithm 1 Workflow of Dis UE |
| Open Source Code | No | The paper does not provide an explicit statement about open-sourcing the code or a link to a code repository for the methodology described. |
| Open Datasets | Yes | We evaluate Dis UE on three standard benchmarks: SVHN [Netzer et al., 2011], CIFAR-10 [Krizhevsky et al., 2009], and CIFAR-100 [Krizhevsky et al., 2009]. |
| Dataset Splits | No | The paper describes partitioning datasets for client heterogeneity using a Dirichlet distribution (Dir(ϵ)), but it does not explicitly provide the training/test/validation dataset splits (e.g., specific percentages or sample counts) for model evaluation. |
| Hardware Specification | No | The paper does not provide specific hardware details such as GPU models, CPU types, or memory specifications used for running the experiments. |
| Software Dependencies | No | The paper does not provide specific software dependencies with version numbers (e.g., Python, PyTorch, TensorFlow versions). |
| Experiment Setup | Yes | For all methods, we set the communication rounds T = 500, the number of clients N = 100, with an active fraction Act = 0.15. For local training, we set the number of local epochs local E = 5, batch size = 50, and the weight decay to 1 10 3. The learning rates for the classifier and generator are initialized to 0.1 and 0.01, respectively. The dimension z is set to 100 for CIFAR-10 and SVHN, and 256 for CIFAR-100. Unless otherwise specified, we adopt βcf = 1.0 and βdiv = 1.0. All our experimental results represent the average over five random seeds. |