Research map / Optimization / Distributed and compressed
Improved Communication Efficiency in Federated Natural Policy Gradient via ADMM-based Gradient Updates
Guangchen Lan, Han Wang, James Anderson, Christopher Brinton, Vaneet Aggarwal
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 |
|---|---|---|---|
| replace standard average FedNPG with ADMM-based direction estimation | FedNPG · Swimmer-v4, 8 agents | final reward: 124.8 -> 128.5 | No clear effect |
| replace standard average FedNPG with ADMM-based direction estimation | FedNPG · Hopper-v4, 8 agents | final reward: 2736 -> 2719 | No clear effect |
| replace standard average FedNPG with ADMM-based direction estimation | FedNPG · Swimmer-v4 and Humanoid-v4 | communication overhead: 4 orders of magnitude (Swimmer), about 6 (Humanoid) lower | Improved |
| increase federated agents from 1 to 8 (ablation) | FedNPG-ADMM · Swimmer-v4 | final reward: 109.4 -> 128.5 | Improved |
| increase federated agents from 1 to 8 (ablation) | FedNPG-ADMM · Hopper-v4 | final reward: 1473 -> 2719 | Improved |
| use second-order FedNPG methods instead of first-order FedPPO | FedNPG-ADMM · Swimmer-v4 and Humanoid-v4 | reward and convergence speed | Improved |
| aggregate only 50% or 75% of agents per iteration (ablation) | FedNPG-ADMM · Swimmer-v4 | final reward: drop less than 6% | Hurt |
Closest papers by meaning
- Asynchronous Federated Reinforcement Learning with Policy Gradient Updates: Algorithm Design and Convergence Analysis (ICLR 2025 · similarity 0.90)
- Federated Natural Policy Gradient and Actor Critic Methods for Multi-task Reinforcement Learning (NeurIPS 2024 · similarity 0.89)
- Momentum for the Win: Collaborative Federated Reinforcement Learning across Heterogeneous Environments (ICML 2024 · similarity 0.88)
- Finite-Time Analysis of On-Policy Heterogeneous Federated Reinforcement Learning (ICLR 2024 · similarity 0.88)
- Federated Offline Reinforcement Learning: Collaborative Single-Policy Coverage Suffices (ICML 2024 · similarity 0.87)
- On the Linear Speedup of Personalized Federated Reinforcement Learning with Shared Representations (ICLR 2025 · similarity 0.87)
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.
