# Test-Time Perturbation Learning with Delayed Feedback for Vision-Language-Action Models

arXiv: https://arxiv.org/abs/2604.18107 · HTML: https://mlsanity.com/paper/2604.18107/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Efficient VLA inference and adaptation, one of 13 directions in Vision-language-action models.
- Builds on 27 of its 33 references that are in the map.
- First posted on arXiv 20 Apr 2026.

## Placed in

- Vision-language-action models / [Efficient VLA inference and adaptation](https://mlsanity.com/vla/efficient-vla-inference-and-adaptation/)

## Builds on (most cited first, 27 of 27)

- [RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control](https://mlsanity.com/paper/2307.15818/) (2023)
- [Diffusion Policy: Visuomotor Policy Learning via Action Diffusion](https://mlsanity.com/paper/2303.04137/) (2023)
- [π0: A Vision-Language-Action Flow Model for General Robot Control](https://mlsanity.com/paper/2410.24164/) (2024)
- [Octo: An Open-Source Generalist Robot Policy](https://mlsanity.com/paper/2405.12213/) (2024)
- [LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learning](https://mlsanity.com/paper/2306.03310/) (2023)
- [PackNet: Adding Multiple Tasks to a Single Network by Iterative Pruning](https://mlsanity.com/paper/1711.05769/) (2017)
- [Fine-Tuning Vision-Language-Action Models: Optimizing Speed and Success](https://mlsanity.com/paper/2502.19645/) (2025)
- [A Generalist Agent](https://mlsanity.com/paper/2205.06175/) (2022)
- [SpatialVLA: Exploring Spatial Representations for Visual-Language-Action Model](https://mlsanity.com/paper/2501.15830/) (2025)
- [CoT-VLA: Visual Chain-of-Thought Reasoning for Vision-Language-Action Models](https://mlsanity.com/paper/2503.22020/) (2025)
- [An Embodied Generalist Agent in 3D World](https://mlsanity.com/paper/2311.12871/) (2023)
- [Vision-Language Foundation Models as Effective Robot Imitators](https://mlsanity.com/paper/2311.01378/) (2023)
- [3D-VLA: A 3D Vision-Language-Action Generative World Model](https://mlsanity.com/paper/2403.09631/) (2024)
- [TraceVLA: Visual Trace Prompting Enhances Spatial-Temporal Awareness for Generalist Robotic Policies](https://mlsanity.com/paper/2412.10345/) (2024)
- [LIBERO-Plus: In-depth Robustness Analysis of Vision-Language-Action Models](https://mlsanity.com/paper/2510.13626/) (2025)
- [Any-point Trajectory Modeling for Policy Learning](https://mlsanity.com/paper/2401.00025/) (2023)
- [SimpleVLA-RL: Scaling VLA Training via Reinforcement Learning](https://mlsanity.com/paper/2509.09674/) (2025)
- [Steering Your Generalists: Improving Robotic Foundation Models via Value Guidance](https://mlsanity.com/paper/2410.13816/) (2024)
- [RoboMonkey: Scaling Test-Time Sampling and Verification for Vision-Language-Action Models](https://mlsanity.com/paper/2506.17811/) (2025)
- [TGRPO :Fine-tuning Vision-Language-Action Model via Trajectory-wise Group Relative Policy Optimization](https://mlsanity.com/paper/2506.08440/) (2025)
- [JARVIS-VLA: Post-Training Large-Scale Vision Language Models to Play Visual Games with Keyboards and Mouse](https://mlsanity.com/paper/2503.16365/) (2025)
- [Verifier-free Test-Time Sampling for Vision-Language-Action Models](https://mlsanity.com/paper/2510.05681/) (2025)
- [Jack of All Trades, Master of Some, a Multi-Purpose Transformer Agent](https://mlsanity.com/paper/2402.09844/) (2024)
- [SCAP: Transductive Test-Time Adaptation via Supportive Clique-based Attribute Prompting](https://mlsanity.com/paper/2503.12866/) (2025)
- [MetaVLA: Unified Meta Co-training For Efficient Embodied Adaption](https://mlsanity.com/paper/2510.05580/) (2025)
- [Test-Time Adaptation with Binary Feedback](https://mlsanity.com/paper/2505.18514/) (2025)
- [Class-aware Domain Knowledge Fusion and Fission for Continual Test-Time Adaptation](https://mlsanity.com/paper/2510.12150/) (2025)

## Abstract (arXiv)

Vision-Language-Action models (VLAs) achieve remarkable performance in sequential decision-making but remain fragile to subtle environmental shifts, such as small changes in object pose. We attribute this brittleness to trajectory overfitting, where VLAs over-attend to the spurious correlation between actions and entities, then reproduce memorized action patterns. We propose Perturbation learning with Delayed Feedback (PDF), a verifier-free test-time adaptation framework that improves decision performance without fine-tuning the base model. PDF mitigates the spurious correlation through uncertainty-based data augmentation and action voting, while an adaptive scheduler allocates augmentation budgets to balance performance and efficiency. To further improve stability, PDF learns a lightweight perturbation module that retrospectively adjusts action logits guided by delayed feedback, correcting overconfidence issue. Experiments on LIBERO (+7.4% success rate) and Atari (+10.3 human normalized score) demonstrate consistent gains of PDF in task success over vanilla VLA and VLA with test-time adaptation, establishing a practical path toward reliable test-time adaptation in multimodal decision-making agents. The code is available at hrefhttps://github.com/zhoujiahuan1991/CVPR2026-PDFhttps://github.com/zhoujiahuan1991/CVPR2026-PDF.