Research map / Reinforcement learning / Policy gradient and actor-critic
Harmonized Dual Policy Improvement for Modelic Reinforcement Learning
Guojian Zhan, Likun Wang, Feihong Zhang, Yang Guan, Shengbo Eben Li
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 |
|---|---|---|---|
| add harmonized dual policy gradient (HDPI) to BOOM | BOOM · DMC Suite (7 tasks) | mean TAR: 803.1 -> 852.6 | Improved |
| add harmonized dual policy gradient (HDPI) to BOOM | BOOM · Humanoid-Bench (7 tasks) | mean TAR: 639.0 -> 746.1 | Improved |
| add harmonized dual policy gradient (HDPI) to BOOM | BOOM · H1-Run | TAR: 330.3 -> 752.1 | Improved |
| add harmonized dual policy gradient (HDPI) to BOOM | BOOM · H1-Walk | TAR: 902.1 -> 929.4 | No clear effect |
| add harmonization step to policy update | BOOM · wall-clock for 10k iterations | seconds: 312 -> 339 | Hurt |
| vary harmonization coefficient rho by 0.1x, 1x, 10x (ablation) | BOOM-H · Dog-run | return and policy loss | No clear effect |
| vary alignment coefficient lambda by 0.1x, 1x, 10x (ablation) | BOOM-H · Dog-run | asymptotic return | Improved |
| replace naive weighted combination with harmonized update (ablation) | BOOM · Dog-run and H1hand-run | policy loss | Improved |
Closest papers by meaning
- Bootstrap Off-policy with World Model (NeurIPS 2025 · similarity 0.88)
- Stabilizing Policy Optimization via Logits Convexity (NeurIPS 2026 · similarity 0.86)
- The Ladder in Chaos: Improving Policy Learning by Harnessing the Parameter Evolving Path in A Low-dimensional Space (NeurIPS 2024 · similarity 0.86)
- Flattening Hierarchies with Policy Bootstrapping (NeurIPS 2025 · similarity 0.86)
- PARL: A Unified Framework for Policy Alignment in Reinforcement Learning from Human Feedback (ICLR 2024 · similarity 0.86)
- Diffusion Policy Policy Optimization (ICLR 2025 · similarity 0.85)
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.
