Research map / Reinforcement learning / Policy gradient and actor-critic
Exploring Safer Behaviors for Deep 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 |
|---|---|---|---|
| Add Safety-Oriented Search to PPO, using safe mutations and Estimated Verification | PPO · Safety Gym tasks | auxiliary cost | Helped |
| Add Safety-Oriented Search to PPO, using safe mutations and Estimated Verification | PPO · Safety Gym tasks | return | No clear effect |
| Replace safe mutations with Gaussian noise and remove the cost-buffer (ablation) | SOS-PPO · PointGoal1 | return | Hurt |
| Replace safe mutations with Gaussian noise and remove the cost-buffer (ablation) | SOS-PPO · PointGoal1 | cost | Hurt |
| Compute mutation sensitivity on arbitrary samples instead of unsafe states (ablation) | SOS-PPO · PointGoal1 | return | Hurt |
| Compute mutation sensitivity on arbitrary samples instead of unsafe states (ablation) | SOS-PPO · PointGoal1 | cost | Hurt |
| Remove Estimated Verification and select the comparable-return individual with minimum cost (ablation) | SOS-PPO · PointGoal1 | return | Hurt |
| Remove Estimated Verification and select the comparable-return individual with minimum cost (ablation) | SOS-PPO · PointGoal1 | cost | Hurt |
| Use Estimated Verification with cost reachability sets instead of estimated reachability sets (ablation) | Estimated Verification · Safety properties | computation time: approximately half the time | Helped |
Closest papers by meaning
- Towards Safe Reinforcement Learning with a Safety Editor Policy (NeurIPS 2022 · similarity 0.92)
- Balance Reward and Safety Optimization for Safe Reinforcement Learning: A Perspective of Gradient Manipulation (AAAI 2024 · similarity 0.92)
- Safety Representations for Safer Policy Learning (ICLR 2025 · similarity 0.91)
- Enhancing Efficiency of Safe Reinforcement Learning via Sample Manipulation (NeurIPS 2024 · similarity 0.90)
- Sampling-Based Safe Reinforcement Learning (NeurIPS 2026 · similarity 0.90)
- Safe Reinforcement Learning with Instantaneous Constraints: The Role of Aggressive Exploration (AAAI 2024 · similarity 0.90)
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.