Pith. sign in

REVIEW 18 cited by

IRL-VLA: Training an Vision-Language-Action Policy via Reward World Model

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2508.06571 v3 pith:47NZQJKU submitted 2025-08-07 cs.AI cs.CVcs.RO

IRL-VLA: Training an Vision-Language-Action Policy via Reward World Model

classification cs.AI cs.CVcs.RO
keywords learningrewardclose-loopdrivingmodelworldautonomousperformance
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

Vision-Language-Action (VLA) models have demonstrated potential in autonomous driving. However, two critical challenges hinder their development: (1) Existing VLA architectures are typically based on imitation learning in open-loop setup which tends to capture the recorded behaviors in the dataset, leading to suboptimal and constrained performance, (2) Close-loop training relies heavily on high-fidelity sensor simulation, where domain gaps and computational inefficiencies pose significant barriers. In this paper, we introduce IRL-VLA, a novel close-loop Reinforcement Learning via \textbf{I}nverse \textbf{R}einforcement \textbf{L}earning reward world model with a self-built VLA approach. Our framework proceeds in a three-stage paradigm: In the first stage, we propose a VLA architecture and pretrain the VLA policy via imitation learning. In the second stage, we construct a lightweight reward world model via inverse reinforcement learning to enable efficient close-loop reward computation. To further enhance planning performance, finally, we design specialized reward world model guidence reinforcement learning via PPO(Proximal Policy Optimization) to effectively balance the safety incidents, comfortable driving, and traffic efficiency. Our approach achieves state-of-the-art performance in NAVSIM v2 end-to-end driving benchmark, 1st runner up in CVPR2025 Autonomous Grand Challenge. We hope that our framework will accelerate VLA research in close-loop autonomous driving.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 18 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Latent Chain-of-Thought World Modeling for End-to-End Driving

    cs.CV 2025-12 unverdicted novelty 7.0

    LCDrive unifies chain-of-thought reasoning and action selection for end-to-end driving by interleaving action-proposal tokens and latent world-model tokens that predict action outcomes, yielding faster inference and b...

  2. WCog-VLA: A Dual-Level World-Cognitive Vision-Language-Action Model for End-to-End Autonomous Driving

    cs.CV 2026-07 conditional novelty 6.0

    WCog-VLA couples Game-CoT semantic reasoning with an aligned decoupled diffusion transformer to generate joint multi-agent trajectories and reaches 92.9 PDMS on NAVSIM.

  3. World Engine: Towards the Era of Post-Training for Autonomous Driving

    cs.RO 2026-06 unverdicted novelty 6.0

    World Engine generates realistic safety-critical driving variations from logs for reinforcement post-training, reducing benchmark failures more than data scaling and showing collision reductions plus on-road gains in ...

  4. IDOL: Inverse-Dynamics-Guided Future Prediction for End-to-End Autonomous Driving

    cs.RO 2026-05 unverdicted novelty 6.0

    IDOL uses inverse dynamics on adjacent predicted latent futures to extract planning-relevant motion deltas, then optimizes trajectories with a closed-loop refinement step, reporting SOTA results on NAVSIM v1 and v2.

  5. NTR: Neural Token Reconstruction for Scene Token Bottleneck in End-to-End Driving

    cs.CV 2026-05 unverdicted novelty 6.0

    NTR adds a self-distillation masked latent reconstruction objective that uses only scene tokens to reconstruct masked patch features, improving visual representation quality and planning performance in end-to-end auto...

  6. Distill to Think, Foresee to Act: Cognitive-Physical Reinforcement Learning for Autonomous Driving

    cs.CV 2026-05 unverdicted novelty 6.0

    CoPhy distills VLM knowledge into a BEV encoder and uses an action-conditioned auto-regressive BEV world model inside GRPO with dual physical-cognitive rewards to reach SOTA on NAVSIM v1/v2 while adding language-based...

  7. Human Cognition in Machines: A Unified Perspective of World Models

    cs.RO 2026-04 unverdicted novelty 6.0

    The paper introduces a unified framework for world models that fully incorporates all cognitive functions from Cognitive Architecture Theory, highlights under-researched areas in motivation and meta-cognition, and pro...

  8. Alpamayo-R1: Bridging Reasoning and Action Prediction for Generalizable Autonomous Driving in the Long Tail

    cs.RO 2025-10 conditional novelty 6.0

    Alpamayo-R1 introduces a VLA model with a Chain of Causation dataset and multi-stage SFT-plus-RL training that reports 12% better planning accuracy and 35% fewer close encounters versus trajectory-only baselines in dr...

  9. World-Env: Leveraging World Model as a Virtual Environment for VLA Post-Training

    cs.RO 2025-09 unverdicted novelty 6.0

    World-Env replaces physical robot interactions with a world model-based virtual environment and VLM-guided rewards to enable efficient RL post-training for VLA models, showing gains with only five demonstrations per task.

  10. Discrete-WAM: Unified Discrete Vision-Action Token Editing for World-Policy Learning

    cs.RO 2026-06 unverdicted novelty 5.0

    Discrete-WAM unifies world modeling and policy learning for autonomous driving by representing observations, states, decisions, and actions as tokens in one space and using hierarchical token editing for planning.

  11. World Models for Robotic Manipulation: A Survey

    cs.RO 2026-05 accept novelty 5.0

    Survey organizing world models for robotic manipulation into representation families, a functional taxonomy, and infrastructure roles across pretraining, post-training, and inference, while reviewing 34 datasets and e...

  12. Distill to Think, Foresee to Act: Cognitive-Physical Reinforcement Learning for Autonomous Driving

    cs.CV 2026-05 unverdicted novelty 5.0

    CoPhy is a new RL framework that distills VLM cognition into BEV encoders, adds an auto-regressive BEV world model for action-conditioned future prediction, and optimizes policies via GRPO with dual physical-cognitive...

  13. Latency Analysis and Optimization of Alpamayo 1 via Efficient Trajectory Generation

    cs.AI 2026-05 unverdicted novelty 5.0

    Redesigning Alpamayo 1 to single-reasoning and optimizing diffusion action generation cuts inference latency by 69.23% while preserving trajectory diversity and prediction quality.

  14. SpanVLA: Efficient Action Bridging and Learning from Negative-Recovery Samples for Vision-Language-Action Model

    cs.CV 2026-04 unverdicted novelty 5.0

    SpanVLA reduces action generation latency via flow-matching conditioned on history and improves robustness by training on negative-recovery samples with GRPO and a dedicated reasoning dataset.

  15. AVA-VLA: Improving Vision-Language-Action models with Active Visual Attention

    cs.LG 2025-11 unverdicted novelty 5.0

    AVA-VLA reformulates VLA learning as a POMDP using recurrent states and active visual attention to achieve state-of-the-art results on LIBERO, CALVIN, and real dual-arm tasks.

  16. Post-Training in End-to-End Autonomous Driving

    cs.CV 2026-07 unverdicted novelty 4.0

    Post-training for end-to-end autonomous driving is surveyed and grouped into four supervision-based families to address limits of open-loop imitation.

  17. Vision-and-Language Navigation for UAVs: Progress, Challenges, and a Research Roadmap

    cs.RO 2026-04 unverdicted novelty 4.0

    A survey of UAV vision-and-language navigation that establishes a methodological taxonomy, reviews resources and challenges, and proposes a forward-looking research roadmap.

  18. Post-Training in End-to-End Autonomous Driving

    cs.CV 2026-07 accept novelty 2.0

    A survey organizing post-training methods for autonomous driving into four supervision-based families: distillation, preference alignment, reinforcement learning, and test-time refinement.