# Safe and constrained RL

Area: [Deep reinforcement learning methods](https://mlsanity.com/deeprl/index.md). HTML: https://mlsanity.com/deeprl/safe-and-constrained-rl/

## Key facts

- Safe and constrained RL: 323 papers on arXiv since 2024 (data through 14 Oct 2026).
- 16 of its 323 mapped papers (5%) cite “A Survey of Constraint Formulations in Safe Reinforcement Learning” (2024, arXiv 2402.02025).
- 12 of its 323 mapped papers (4%) cite “Balance Reward and Safety Optimization for Safe Reinforcement Learning: A Perspective of Gradient Manipulation” (2024, arXiv 2405.01677).
- 10 new papers 30 Sep–14 Oct.

## Most cited here

- [A Survey of Constraint Formulations in Safe Reinforcement Learning](https://mlsanity.com/paper/2402.02025/) (2024), cited by 16 papers here
- [Balance Reward and Safety Optimization for Safe Reinforcement Learning: A Perspective of Gradient Manipulation](https://mlsanity.com/paper/2405.01677/) (2024), cited by 12 papers here
- [ActSafe: Active Exploration with Safety Constraints for Reinforcement Learning](https://mlsanity.com/paper/2410.09486/) (2024), cited by 11 papers here
- [Dynamic Model Predictive Shielding for Provably Safe Reinforcement Learning](https://mlsanity.com/paper/2405.13863/) (2024), cited by 7 papers here
- [Safety Filtering While Training: Improving the Performance and Sample Efficiency of Reinforcement Learning Agents](https://mlsanity.com/paper/2410.11671/) (2024), cited by 7 papers here
- [SPiDR: A Simple Approach for Zero-Shot Safety in Sim-to-Real Transfer](https://mlsanity.com/paper/2509.18648/) (2025), cited by 6 papers here
- [Off-Policy Primal-Dual Safe Reinforcement Learning](https://mlsanity.com/paper/2401.14758/) (2024), cited by 5 papers here
- [Reinforcement Learning-based Receding Horizon Control using Adaptive Control Barrier Functions for Safety-Critical Systems](https://mlsanity.com/paper/2403.17338/) (2024), cited by 5 papers here
- [Safe Reinforcement Learning on the Constraint Manifold: Theory and Applications](https://mlsanity.com/paper/2404.09080/) (2024), cited by 5 papers here
- [Enhancing Efficiency of Safe Reinforcement Learning via Sample Manipulation](https://mlsanity.com/paper/2405.20860/) (2024), cited by 5 papers here

## New (30 Sep–14 Oct)

- [Safe Reinforcement Learning with Receding-Horizon Safety Critic and Differentiable Action Shields](https://mlsanity.com/paper/2610.20255/) (2026-10-14)
- [Certifying Almost-Sure Reachability in Unknown MDPs](https://mlsanity.com/paper/2610.18119/) (2026-10-13)
- [FAITH: Feasibility-Aware Safety-Filtered RL for High-Dimensional Systems](https://mlsanity.com/paper/2610.12432/) (2026-10-08)
- [A Unified Bellman Operator for Safety-Critical Reinforcement Learning](https://mlsanity.com/paper/2610.12420/) (2026-10-08)
- [Higher-Order Action Supervision Makes A Strong Policy Class](https://mlsanity.com/paper/2610.11175/) (2026-10-08)
- [Safe on Average, Unsafe in the Tail: When Is the Episodic-Cost Tail Controllable?](https://mlsanity.com/paper/2610.09508/) (2026-10-07)
- [Reachability-Aware Diffusion Policy Optimization](https://mlsanity.com/paper/2610.05969/) (2026-10-05)
- [FoSeRL: Formal Sequential Robustness Certification for Reinforcement Learning Policies](https://mlsanity.com/paper/2610.04754/) (2026-10-03)
- [CORE-RL: Confidence-Oriented Reliability Evaluation of Black-Box Reinforcement Learning Policies](https://mlsanity.com/paper/2610.04418/) (2026-10-03)
- [DODGER: Safety-Guided Reinforcement Learning for Robot Navigation Among Dynamic Obstacles](https://mlsanity.com/paper/2609.38873/) (2026-09-30)