AI

Fair Federated Learning on Heterogeneous Data
We discuss why previous fairness metrics fail when clients have heterogeneous data contribution, and design new metrics for fair Federated Learning for clients with heterogeneous data. Clustering is guaranteed to help improve fairness for Federated Learning.
PDF arxiv
Fair Federated Learning on Heterogeneous Data
Diversifying Options in Option-Critic Framework of Hierarchical Reinforcement Learning
Implement the Option-Critic architecture and diversify its choice of options.
PDF
Diversifying Options in Option-Critic Framework of Hierarchical Reinforcement Learning