AI Paper Atlas

Research map / Reinforcement learning / Visual and representation RL

Unleashing VLA Potentials in Autonomous Driving via Explicit Learning from Failures

Yuechen Luo, Fang Li, Qimao Chen, Shaoqing Xu, Jiaxin Liu, Ziying Song, Zhi-xin Yang, Fuxi Wen

CVPR 2026 · Reinforcement learning › Visual and representation RL · approach: Probabilistic and autonomous RL · 13 citations (OpenAlex)

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 changedWhereMeasuredEffect
add teacher-feedback-guided refinement samples to GRPOGRPO · NAVSIMv1PDMS: 89.0 -> 91.0Improved
add ground-truth trajectories directly to GRPO batch (ablation)GRPO · NAVSIMv1PDMS: 89.0 -> 89.2Improved
replace rule-based feedback with teacher structured feedback (ablation)Rule-GRPO · NAVSIMv1PDMS: 89.6 -> 91.0Improved
replace GT trajectory injection with feedback-guided refinement (ablation)GT-GRPO · NAVSIMv1PDMS: 89.2 -> 91.0Improved
add teacher feedback refinementGRPO · NAVSIMv1 high-level planningaccuracy (%): 79.3 -> 80.3Improved
add teacher feedback refinementGRPO · RL total-failure ratioPDMS total-failure rate (%): 2.73 -> 1.08Improved
curate difficult and ambiguous 24k samples instead of full 85k (ablation)ELF-VLA · NAVSIMv1PDMS: 89.1 -> 91.0Improved
curate 24k instead of random 24k (ablation)ELF-VLA · NAVSIMv1PDMS: 88.9 -> 91.0Improved
use k=1 refinement response instead of k=4 (ablation)ELF-VLA · NAVSIMv1PDMS: 89.0 -> 91.0Improved
use k=1 instead of k=2 refinement responses (ablation)ELF-VLA · NAVSIMv1PDMS: 89.7 -> 91.0Improved
add policy shaping on feedback samples (ablation)ELF-VLA · NAVSIMv1PDMS: 89.3 -> 91.0Improved
use ELF-VLA vs DriveSuprimDriveSuprim · NAVSIMv2EPDMS: 87.1 -> 87.1No clear effect

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy