Research map / Reinforcement learning / Policy gradient and actor-critic
Neurosymbolic Reinforcement Learning with Formally Verified Exploration
Greg Anderson, Abhinav Verma, Isil Dillig, Swarat Chaudhuri
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 |
|---|---|---|---|
| update the shield during learning instead of keeping a static shield (ablation) | REVEL · 10 continuous-control benchmarks | training cost: better on 4 of 10 benchmarks, very substantial on noisy-road and mountain-car | Improved |
| use REVEL's neurosymbolic verified policy instead of DDPG | REVEL · 10 continuous-control benchmarks | policy cost: competitive or better on 7 of 10; slightly worse on 2; much better on car-racing | Mixed |
| use REVEL instead of CPO | REVEL · 10 continuous-control benchmarks | policy cost: better on 4 of 10, slightly worse on 2 | Mixed |
| enforce verified shielding (REVEL) vs unverified DDPG/CPO | REVEL · 10 benchmarks | safety violations during training: DDPG and CPO violate safety in 8 of 10 benchmarks, e.g. CPO 9906 on pendulum | Improved |
| repeated shield synthesis inside the learning loop | REVEL · 10 benchmarks | training time: 87% of time in shield synthesis on average; 95.1% on road-2d | Hurt |
Closest papers by meaning
- Verified Safe Reinforcement Learning for Neural Network Dynamic Models (NeurIPS 2024 · similarity 0.88)
- Sampling-Based Safe Reinforcement Learning (NeurIPS 2026 · similarity 0.87)
- BlendRL: A Framework for Merging Symbolic and Neural Policy Learning (ICLR 2025 · similarity 0.87)
- Safe Reinforcement Learning with Instantaneous Constraints: The Role of Aggressive Exploration (AAAI 2024 · similarity 0.87)
- ActSafe: Active Exploration with Safety Constraints for Reinforcement Learning (ICLR 2025 · similarity 0.86)
- Neurosymbolic Reinforcement Learning: Playing MiniHack with Probabilistic Logic Shields (AAAI 2025 · similarity 0.86)
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.
