# ADReFT: Adaptive Decision Repair for Safe Autonomous Driving via Reinforcement Fine-Tuning

arXiv: https://arxiv.org/abs/2506.23960 · HTML: https://mlsanity.com/paper/2506.23960/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Reinforcement learning for driving, one of 9 directions in End-to-end autonomous driving.
- Builds on 15 of its 57 references that are in the map.
- First posted on arXiv 30 Jun 2025.

## Placed in

- End-to-end autonomous driving / [Reinforcement learning for driving](https://mlsanity.com/driving/reinforcement-learning-for-driving/)

## Builds on (most cited first, 15 of 15)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [BEVFusion: Multi-Task Multi-Sensor Fusion with Unified Bird's-Eye View Representation](https://mlsanity.com/paper/2205.13542/) (2022)
- [End-to-End Urban Driving by Imitating a Reinforcement Learning Coach](https://mlsanity.com/paper/2108.08265/) (2021)
- [PlanT: Explainable Planning Transformers via Object-Level Representations](https://mlsanity.com/paper/2210.14222/) (2022)
- [SafeBench: A Benchmarking Platform for Safety Evaluation of Autonomous Vehicles](https://mlsanity.com/paper/2206.09682/) (2022)
- [Deep Imitation Learning for Autonomous Driving in Generic Urban Scenarios with Enhanced Safety](https://mlsanity.com/paper/1903.00640/) (2019)
- [Pylot: A Modular Platform for Exploring Latency-Accuracy Tradeoffs in Autonomous Vehicles](https://mlsanity.com/paper/2104.07830/) (2021)
- [Neural Network Guided Evolutionary Fuzzing for Finding Traffic Violations of Autonomous Vehicles](https://mlsanity.com/paper/2109.06126/) (2021)
- [Many-Objective Reinforcement Learning for Online Testing of DNN-Enabled Systems](https://mlsanity.com/paper/2210.15432/) (2022)
- [DriveFuzz: Discovering Autonomous Driving Bugs through Driving Quality-Guided Fuzzing](https://mlsanity.com/paper/2211.01829/) (2022)
- [LawBreaker: An Approach for Specifying Traffic Laws and Fuzzing Autonomous Vehicles](https://mlsanity.com/paper/2208.14656/) (2022)
- [REDriver: Runtime Enforcement for Autonomous Vehicles](https://mlsanity.com/paper/2401.02253/) (2024)
- [FixDrive: Automatically Repairing Autonomous Vehicle Driving Behaviour for 0.08 per Violation](https://mlsanity.com/paper/2502.08260/) (2025)

## Built on by (most cited first, 1 of 1)

- [SafeAlign-VLA: A Negative-Enhanced Safe Alignment Framework for Risk-Aware Autonomous Driving](https://mlsanity.com/paper/2605.19524/) (2026)

## Abstract (arXiv)

Autonomous Driving Systems (ADSs) continue to face safety-critical risks due to the inherent limitations in their design and performance capabilities. Online repair plays a crucial role in mitigating such limitations, ensuring the runtime safety and reliability of ADSs. Existing online repair solutions enforce ADS compliance by transforming unacceptable trajectories into acceptable ones based on predefined specifications, such as rule-based constraints or training datasets. However, these approaches often lack generalizability, adaptability and tend to be overly conservative, resulting in ineffective repairs that not only fail to mitigate safety risks sufficiently but also degrade the overall driving experience. To address this issue, we propose Adaptive Decision Repair (ADReFT), a novel and effective repair method that identifies safety-critical states through offline learning from failed tests and generates appropriate mitigation actions to improve ADS safety. Specifically, ADReFT incorporates a transformer-based model with two joint heads, State Monitor and Decision Adapter, designed to capture complex driving environment interactions to evaluate state safety severity and generate adaptive repair actions. Given the absence of oracles for state safety identification, we first pretrain ADReFT using supervised learning with coarse annotations, i.e., labeling states preceding violations as positive samples and others as negative samples. It establishes ADReFT's foundational capability to mitigate safety-critical violations, though it may result in somewhat conservative mitigation strategies. Therefore, we subsequently finetune ADReFT using reinforcement learning to improve its initial capability and generate more precise and contextually appropriate repair decisions. Our evaluation results illustrate that ADReFT achieves better repair performance.