Research map / Reinforcement learning / Policy gradient and actor-critic
Augmented Proximal Policy Optimization for Safe Reinforcement Learning
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 P3O instead of CPO | Safe RL policy optimization · AntCircle | Reward: +15% | Helped |
| Use P3O instead of PPO-Lagrangian | Safe RL policy optimization · AntCircle | Reward: +24% | Helped |
| Use P3O instead of FOCOPS | Safe RL policy optimization · AntCircle | Reward: +9% | Helped |
| Use P3O instead of CPO | Safe RL policy optimization · PointGather | Reward: +34% | Helped |
| Use P3O instead of PPO-Lagrangian | Safe RL policy optimization · PointGather | Reward: +23% | Helped |
| Use P3O instead of FOCOPS | Safe RL policy optimization · PointGather | Reward: +31% | Helped |
| Include both hazard and pillar penalties rather than only one (ablation) | P3O · Navigation | constraint satisfaction | Helped |
| Add a centralized critic using concatenated local observations (ablation) | MAP3O · Simple Spread | Reward | Helped |
| Normalize advantage estimates to use a fixed penalty factor | P3O · Safe RL tasks | stability across penalty settings | Helped |
Closest papers by meaning
- P2BPO: Permeable Penalty Barrier-Based Policy Optimization for Safe RL (AAAI 2024 · similarity 0.90)
- Off-Policy Primal-Dual Safe Reinforcement Learning (ICLR 2024 · similarity 0.90)
- Enhancing Efficiency of Safe Reinforcement Learning via Sample Manipulation (NeurIPS 2024 · similarity 0.90)
- Safe Reinforcement Learning with Instantaneous Constraints: The Role of Aggressive Exploration (AAAI 2024 · similarity 0.89)
- Reward Penalties on Augmented States for Solving Richly Constrained RL Effectively (AAAI 2024 · similarity 0.89)
- Enhancing Safety in Reinforcement Learning with Human Feedback via Rectified Policy Optimization (NeurIPS 2025 · similarity 0.88)
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.