Research map / Reinforcement learning / Policy gradient and actor-critic
CEM: Constrained Entropy Maximization for Task-Agnostic Safe Exploration
Qisong Yang, Matthijs T.J. Spaan
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 |
|---|---|---|---|
| Use CEM instead of MEPOL for task-agnostic exploration | Task-agnostic exploration policy · BasicNav, MountainCar, CartPole, PointGoal, CarButton | safety constraint satisfaction and state entropy | Mixed |
| Use CEM instead of SAC-λ-RF | Task-agnostic exploration policy · BasicNav, MountainCar, CartPole, PointGoal, CarButton | safe state-space exploration | Improved |
| Use CEM instead of SAC-λ-IR | Task-agnostic exploration policy · BasicNav, MountainCar, CartPole, PointGoal, CarButton | safe state-space exploration | Improved |
| Use an unsafe, broadly exploring guide policy | SaGui target-task learning · PointGoal, CarButton | target-task reward | Improved |
| Use an unsafe, broadly exploring guide policy | SaGui target-task learning · PointGoal, CarButton | safety | No clear effect |
| Use a safe but conservative guide policy | SaGui target-task learning · PointGoal, CarButton | target-task reward | Hurt |
| Use a CEM guide policy balancing safety and exploration | SaGui target-task learning · PointGoal, CarButton | target-task reward and safety | Improved |
Closest papers by meaning
- Global Optimality for Constrained Exploration via Penalty Regularization (NeurIPS 2026 · similarity 0.88)
- Safety Representations for Safer Policy Learning (ICLR 2025 · similarity 0.87)
- Absolute State-wise Constrained Policy Optimization: High-Probability State-wise Constraints Satisfaction (NeurIPS 2026 · similarity 0.87)
- SEMDICE: Off-policy State Entropy Maximization via Stationary Distribution Correction Estimation (ICLR 2025 · similarity 0.87)
- Sampling-Based Safe Reinforcement Learning (NeurIPS 2026 · similarity 0.87)
- When Maximum Entropy Misleads Policy Optimization (ICML 2025 · similarity 0.87)
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.
