Unleashing VLA Potentials in Autonomous Driving via Explicit Learning from Failures
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908859048656896 |
|---|---|
| author | Luo, Yuechen Chen, Qimao Li, Fang Xu, Shaoqing Liu, Jaxin Song, Ziying Yang, Zhi-xin Wen, Fuxi |
| author_facet | Luo, Yuechen Chen, Qimao Li, Fang Xu, Shaoqing Liu, Jaxin Song, Ziying Yang, Zhi-xin Wen, Fuxi |
| contents | Vision-Language-Action (VLA) models for autonomous driving often hit a performance plateau during Reinforcement Learning (RL) optimization. This stagnation arises from exploration capabilities constrained by previous Supervised Fine-Tuning (SFT), leading to persistent failures in long-tail scenarios. In these critical situations, all explored actions yield a zero-value driving score. This information-sparse reward signals a failure, yet fails to identify its root cause -- whether it is due to incorrect planning, flawed reasoning, or poor trajectory execution. To address this limitation, we propose VLA with Explicit Learning from Failures (ELF-VLA), a framework that augments RL with structured diagnostic feedback. Instead of relying on a vague scalar reward, our method produces detailed, interpretable reports that identify the specific failure mode. The VLA policy then leverages this explicit feedback to generate a Feedback-Guided Refinement. By injecting these corrected, high-reward samples back into the RL training batch, our approach provides a targeted gradient, which enables the policy to solve critical scenarios that unguided exploration cannot. Extensive experiments demonstrate that our method unlocks the latent capabilities of VLA models, achieving state-of-the-art (SOTA) performance on the public NAVSIM benchmark for overall PDMS, EPDMS score and high-level planning accuracy. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_01063 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Unleashing VLA Potentials in Autonomous Driving via Explicit Learning from Failures Luo, Yuechen Chen, Qimao Li, Fang Xu, Shaoqing Liu, Jaxin Song, Ziying Yang, Zhi-xin Wen, Fuxi Computer Vision and Pattern Recognition Vision-Language-Action (VLA) models for autonomous driving often hit a performance plateau during Reinforcement Learning (RL) optimization. This stagnation arises from exploration capabilities constrained by previous Supervised Fine-Tuning (SFT), leading to persistent failures in long-tail scenarios. In these critical situations, all explored actions yield a zero-value driving score. This information-sparse reward signals a failure, yet fails to identify its root cause -- whether it is due to incorrect planning, flawed reasoning, or poor trajectory execution. To address this limitation, we propose VLA with Explicit Learning from Failures (ELF-VLA), a framework that augments RL with structured diagnostic feedback. Instead of relying on a vague scalar reward, our method produces detailed, interpretable reports that identify the specific failure mode. The VLA policy then leverages this explicit feedback to generate a Feedback-Guided Refinement. By injecting these corrected, high-reward samples back into the RL training batch, our approach provides a targeted gradient, which enables the policy to solve critical scenarios that unguided exploration cannot. Extensive experiments demonstrate that our method unlocks the latent capabilities of VLA models, achieving state-of-the-art (SOTA) performance on the public NAVSIM benchmark for overall PDMS, EPDMS score and high-level planning accuracy. |
| title | Unleashing VLA Potentials in Autonomous Driving via Explicit Learning from Failures |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2603.01063 |