SPACE: Single-round Participant Amalgamation for Contribution Evaluation in Federated Learning

Authors: Yi-Chung Chen, Hsi-Wen Chen, Shun-Gui Wang, Ming-syan Chen

NeurIPS 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results demonstrate that SPACE outperforms state-of-the-art methods in terms of both running time and Pearson s Correlation Coefficient (PCC). Furthermore, extensive experiments conducted on applications, client reweighting, and client selection highlight the effectiveness of SPACE.
Researcher Affiliation Academia Yi-Chung Chen National Taiwan University r10942081@ntu.edu.tw Hsi-Wen Chen National Taiwan University hwchen@arbor.ee.ntu.edu.tw Shun-Gui Wang National Taiwan University r11921099@ntu.edu.tw Ming-Syan Chen National Taiwan University mschen@ntu.edu.tw
Pseudocode Yes Algorithm 1 SPACE
Open Source Code Yes The code is available at https://github.com/culiver/SPACE.
Open Datasets Yes we conduct experiments on the widely adopted image dataset MNIST [24] and CIFAR10 [22].
Dataset Splits No The paper mentions using a 'validation set' multiple times (e.g., 'the evaluation of model performance depends on the size of the validation set on the server.', 'prototypes are built for the server s validation set.'), and uses standard datasets like MNIST and CIFAR10, but does not explicitly provide the specific percentages or counts for training, validation, and test splits.
Hardware Specification Yes Execution time (in seconds) is measured on a single V100 GPU without parallel training to assess the time efficiency of the approaches.
Software Dependencies No The paper does not provide specific software dependencies with version numbers (e.g., Python, PyTorch, TensorFlow versions).
Experiment Setup Yes For adjusting the utility function, we empirically set k as 100 while T as 0.95 and 0.5 for evaluation on MNIST and CIFAR10.