# Automating the Refinement of Reinforcement Learning Specifications

arXiv: https://arxiv.org/abs/2512.01047 · HTML: https://mlsanity.com/paper/2512.01047/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Interpretable and logic-guided RL, one of 15 directions in Deep reinforcement learning methods.
- Builds on 23 of its 37 references that are in the map.
- First posted on arXiv 30 Nov 2025.

## Placed in

- Deep reinforcement learning methods / [Interpretable and logic-guided RL](https://mlsanity.com/deeprl/interpretable-and-logic-guided-rl/)

## Builds on (most cited first, 23 of 23)

- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [End-to-End Training of Deep Visuomotor Policies](https://mlsanity.com/paper/1504.00702/) (2015)
- [Reinforcement Learning With Temporal Logic Rewards](https://mlsanity.com/paper/1612.03471/) (2016)
- [Reinforcement Learning for Temporal Logic Control Synthesis with Probabilistic Satisfaction Guarantees](https://mlsanity.com/paper/1909.05304/) (2019)
- [Omega-Regular Objectives in Model-Free Reinforcement Learning](https://mlsanity.com/paper/1810.00950/) (2018)
- [LTL2Action: Generalizing LTL Instructions for Multi-Task RL](https://mlsanity.com/paper/2102.06858/) (2021)
- [Foundations for Restraining Bolts: Reinforcement Learning with LTLf/LDLf restraining specifications](https://mlsanity.com/paper/1807.06333/) (2018)
- [On the Expressivity of Markov Reward](https://mlsanity.com/paper/2111.00876/) (2021)
- [A Composable Specification Language for Reinforcement Learning Tasks](https://mlsanity.com/paper/2008.09293/) (2020)
- [Q-Learning for Robust Satisfaction of Signal Temporal Logic Specifications](https://mlsanity.com/paper/1609.07409/) (2016)
- [Compositional Reinforcement Learning from Logical Specifications](https://mlsanity.com/paper/2106.13906/) (2021)
- [panda-gym: Open-source goal-conditioned environments for robotic learning](https://mlsanity.com/paper/2106.13687/) (2021)
- [Comprehensive Overview of Reward Engineering and Shaping in Advancing Reinforcement Learning Applications](https://mlsanity.com/paper/2408.10215/) (2024)
- [Transfer of Temporal Logic Formulas in Reinforcement Learning](https://mlsanity.com/paper/1909.04256/) (2019)
- [Stability Verification in Stochastic Control Systems via Neural Network Supermartingales](https://mlsanity.com/paper/2112.09495/) (2021)
- [Learning Control Policies for Stochastic Systems with Reach-avoid Guarantees](https://mlsanity.com/paper/2210.05308/) (2022)
- [DeepLTL: Learning to Efficiently Satisfy Complex LTL Specifications for Multi-Task RL](https://mlsanity.com/paper/2410.04631/) (2024)
- [Compositional Policy Learning in Stochastic Control Systems with Formal Guarantees](https://mlsanity.com/paper/2312.01456/) (2023)
- [Verifiable and Compositional Reinforcement Learning Systems](https://mlsanity.com/paper/2106.05864/) (2021)
- [LCRL: Certified Policy Synthesis via Logically-Constrained Reinforcement Learning](https://mlsanity.com/paper/2209.10341/) (2022)
- [Verifiable Reinforcement Learning Systems via Compositionality](https://mlsanity.com/paper/2309.06420/) (2023)
- [Logical Specifications-guided Dynamic Task Sampling for Reinforcement Learning Agents](https://mlsanity.com/paper/2402.03678/) (2024)
- [On the Hardness of Almost-Sure Termination](https://mlsanity.com/paper/1506.01930/) (2015)

## Abstract (arXiv)

Logical specifications have been shown to help reinforcement learning algorithms in achieving complex tasks. However, when a task is under-specified, agents might fail to learn useful policies. In this work, we explore the possibility of improving coarse-grained logical specifications via an exploration-guided strategy. We propose AutoSpec, a framework that searches for a logical specification refinement whose satisfaction implies satisfaction of the original specification, but which provides additional guidance therefore making it easier for reinforcement learning algorithms to learn useful policies. AutoSpec is applicable to reinforcement learning tasks specified via the SpectRL specification logic. We exploit the compositional nature of specifications written in SpectRL, and design four refinement procedures that modify the abstract graph of the specification by either refining its existing edge specifications or by introducing new edge specifications. We prove that all four procedures maintain specification soundness, i.e. any trajectory satisfying the refined specification also satisfies the original. We then show how AutoSpec can be integrated with existing reinforcement learning algorithms for learning policies from logical specifications. Our experiments demonstrate that AutoSpec yields promising improvements in terms of the complexity of control tasks that can be solved, when refined logical specifications produced by AutoSpec are utilized.