Research map / Multi-agent systems and games / Games and equilibria
Gradient Driven Rewards to Guarantee Fairness in Collaborative Machine Learning
Xinyi Xu, Lingjuan Lyu, Xingjun Ma, Chenglin Miao, Chuan-Sheng Foo, Bryan Kian Hsiang Low
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 |
|---|---|---|---|
| Use cosine-gradient-Shapley-based sparsified rewards instead of FedAvg | federated learning · MNIST 10 agents CLA | average test accuracy (%): 53 -> 75 | Improved |
| Use cosine similarity instead of Euclidean distance for contribution (ablation) | gradient reward mechanism · MNIST 10 agents CLA | average test accuracy (%): 54 -> 75 | Improved |
| Use cosine instead of Euclidean distance (ablation) | gradient reward mechanism · MNIST 10 agents CLA | fairness rho: 94.54 vs 99.82 (EU higher) | Hurt |
| Replace FedAvg with fair gradient reward | federated learning · MNIST 10 agents UNI | fairness rho: -45.60 -> 94.03 | Improved |
| Replace ECI with fair gradient reward | validation-based Shapley · MNIST 10 agents CLA | fairness rho: 99.98 -> 94.54 | Hurt |
| Replace ECI with fair gradient reward | validation-based Shapley · MNIST 10 agents | time overhead (s): 11937.80 -> 4.94 | Improved |
| Increase altruism beta from 1 to 2 (ablation) | fair gradient reward · CIFAR-10 10 agents CLA | average test accuracy (%): 35 -> 36 | Improved |
| Increase altruism beta (ablation) | fair gradient reward · MNIST 10 agents CLA | average test accuracy (%): 74 -> 75 -> 73 | No clear effect |
| Increase altruism beta (ablation) | fair gradient reward · training loss CLA MNIST | training loss | Improved |
| Use fair gradient reward instead of RR | federated learning · MR 5 agents POW | average test accuracy (%): 63 -> 62 | Hurt |
Closest papers by meaning
- Towards Fair Graph Federated Learning via Incentive Mechanisms (AAAI 2024 · similarity 0.88)
- You Get What You Give: Reciprocally Fair Federated Learning (ICML 2025 · similarity 0.88)
- Aequa: Fair Model Rewards in Collaborative Learning via Slimmable Networks (ICML 2025 · similarity 0.87)
- Fair and Efficient Contribution Valuation for Vertical Federated Learning (ICLR 2024 · similarity 0.87)
- Incentivizing Time-Aware Fairness in Data Sharing (NeurIPS 2025 · similarity 0.86)
- Fair Resource Allocation in Multi-Task Learning (ICML 2024 · similarity 0.86)
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.
