Research map / Reinforcement learning / Policy gradient and actor-critic
Safe Reinforcement Learning via Formal Methods: Toward Safe Control Through Proof and 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 |
|---|---|---|---|
| sandbox Q-learning exploration with verified model and controller monitors (JSC) | Q-learning · ACC, accurate model | crashing initial states: 10492 -> 0 | Helped |
| JSC monitors | Q-learning · ACC, accurate model | steady initial states: 72 -> 473 | Helped |
| JSC under 5% model perturbation | Q-learning · ACC, perturbed | crashing initial states: 10561 -> 5 | Helped |
| speculate toward modeled state space (JSCQ) instead of free speculation (ablation) | JSC · ACC, 50% perturbation | crashing initial states: 41 -> 34 | Helped |
| JSCQ at small perturbation (ablation) | JSC · ACC, 5% perturbation | crashing initial states: 3 -> 2 | No clear effect |
Closest papers by meaning
- Verified Safe Reinforcement Learning for Neural Network Dynamic Models (NeurIPS 2024 · similarity 0.90)
- A Physics-Informed Machine Learning Framework for Safe and Optimal Control of Autonomous Systems (ICML 2025 · similarity 0.90)
- Safeguarded Progress in Reinforcement Learning: Safe Bayesian Exploration for Control Policy Synthesis (AAAI 2024 · similarity 0.89)
- Sampling-Based Safe Reinforcement Learning (NeurIPS 2026 · similarity 0.89)
- Probabilistic Shielding for Safe Reinforcement Learning (AAAI 2025 · similarity 0.89)
- Safe Reinforcement Learning with Instantaneous Constraints: The Role of Aggressive Exploration (AAAI 2024 · 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.