Trade-off between Payoff and Model Rewards in Shapley-Fair Collaborative Machine Learning

Authors: Quoc Phong Nguyen, Bryan Kian Hsiang Low, Patrick Jaillet

NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In Sec. 5, we empirically demonstrate sensible characteristics of our proposed allocation scheme in several ML problems. This section empirically illustrates the proposed allocation scheme of both payoff and model rewards in collaborative ML.
Researcher Affiliation Academia Institute of Data Science, National University of Singapore, Republic of Singapore Dept. of Computer Science, National University of Singapore, Republic of Singapore Dept. of Electrical Engineering and Computer Science, MIT, USA
Pseudocode No The paper does not contain any structured pseudocode or algorithm blocks.
Open Source Code Yes https: //github.com/qphong/model-payoff-allocation
Open Datasets Yes To easily interpret the results, we choose the MNIST dataset [10] which contains 70, 000 images of handwritten digits. Additional experiments on the CIFAR-10 [9] dataset and the IMDB movie reviews dataset [11] dataset are in App. J.
Dataset Splits No In the experiments (in Sec. 5), the model performance is measured by the prediction accuracy of models on a validation set which is common across different parties. (The paper mentions a validation set but does not provide specific details about its size or split percentage.)
Hardware Specification No The paper states 'Yes' to the question 'Did you include the total amount of compute and the type of resources used?' in its self-assessment checklist. However, the provided text of the paper does not contain explicit details about the specific hardware (e.g., GPU/CPU models, memory) used for the experiments.
Software Dependencies No The paper does not provide specific software dependencies with version numbers in the main text.
Experiment Setup No The paper describes the dataset partitioning and the use of a validation set but does not provide specific experimental setup details such as hyperparameter values (e.g., learning rate, batch size, epochs) or optimizer settings in the main text.