# STRONG-VLA: Decoupled Robustness Learning for Vision-Language-Action Models under Multimodal Perturbations

arXiv: https://arxiv.org/abs/2604.10055 · HTML: https://mlsanity.com/paper/2604.10055/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- Placed in Safety and adversarial attacks, one of 13 directions in Vision-language-action models.
- Builds on 12 of its 19 references that are in the map.
- First posted on arXiv 11 Apr 2026.

## Placed in

- Vision-language-action models / [Safety and adversarial attacks](https://mlsanity.com/vla/safety-and-adversarial-attacks/)

## Builds on (most cited first, 12 of 12)

- [Towards Deep Learning Models Resistant to Adversarial Attacks](https://mlsanity.com/paper/1706.06083/) (2017)
- [RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control](https://mlsanity.com/paper/2307.15818/) (2023)
- [OpenVLA: An Open-Source Vision-Language-Action Model](https://mlsanity.com/paper/2406.09246/) (2024)
- [π0: A Vision-Language-Action Flow Model for General Robot Control](https://mlsanity.com/paper/2410.24164/) (2024)
- [LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learning](https://mlsanity.com/paper/2306.03310/) (2023)
- [Exploring the Adversarial Vulnerabilities of Vision-Language-Action Models in Robotics](https://mlsanity.com/paper/2411.13587/) (2024)
- [Pure Vision Language Action (VLA) Models: A Comprehensive Survey](https://mlsanity.com/paper/2509.19012/) (2025)
- [VLATest: Testing and Evaluating Vision-Language-Action Models for Robotic Manipulation](https://mlsanity.com/paper/2409.12894/) (2024)
- [Run-time Observation Interventions Make Vision-Language-Action Models More Visually Robust](https://mlsanity.com/paper/2410.01971/) (2024)
- [GEVRM: Goal-Expressive Video Generation Model For Robust Visual Manipulation](https://mlsanity.com/paper/2502.09268/) (2025)
- [Eva-VLA: Evaluating Vision-Language-Action Models' Robustness Under Real-World Physical Variations](https://mlsanity.com/paper/2509.18953/) (2025)
- [When Alignment Fails: Multimodal Adversarial Attacks on Vision-Language-Action Models](https://mlsanity.com/paper/2511.16203/) (2025)

## Built on by (most cited first, 10 of 10)

- [OA-WAM: Object-Addressable World Action Model for Robust Robot Manipulation](https://mlsanity.com/paper/2605.06481/) (2026)
- [Sequential Planning via Anchored Robotic Keypoints](https://mlsanity.com/paper/2606.30613/) (2026)
- [Think Only When Needed: Prompt-Authority Control for Selective Slow-Path Intervention in Vision-Language-Action Manipulation](https://mlsanity.com/paper/2608.23224/) (2026)
- [EATR-Stereo: Embodiment-Aware Token Routing of Paired Stereo Evidence for Humanoid Vision-Language-Action Control](https://mlsanity.com/paper/2608.17453/) (2026)
- [CertVLA: Certified Defense against Physical Visual Attacks for Vision-Language-Action Models](https://mlsanity.com/paper/2608.20791/) (2026)
- [DS-VLA: A Dendritic-inspired Vision-Language-Action Model for Robust Action Control](https://mlsanity.com/paper/2609.32253/) (2026)
- [Are Vision-Language-Action Models Robust to One-Step Observation Perturbations?](https://mlsanity.com/paper/2609.32550/) (2026)
- [RawVLA: Embodied Neural Image Signal Processor For Robotic Manipulation](https://mlsanity.com/paper/2609.37530/) (2026)
- [How (and How Not) to Use Data Augmentation in VLA Post-Training](https://mlsanity.com/paper/2610.05994/) (2026)
- [Adapting Vision-Language-Action Models to Unknown Visual Disruptions During Execution](https://mlsanity.com/paper/2610.07946/) (2026)

## Abstract (arXiv)

Despite their strong performance in embodied tasks, recent Vision-Language-Action (VLA) models remain highly fragile under multimodal perturbations, where visual corruption and linguistic noise jointly induce distribution shifts that degrade task-level execution. Existing robustness approaches typically rely on joint training with perturbed data, treating robustness as a static objective, which leads to conflicting optimization between robustness and task fidelity. In this work, we propose STRONG-VLA, a decoupled fine-tuning framework that explicitly separates robustness acquisition from task-aligned refinement. In Stage I, the model is exposed to a curriculum of multimodal perturbations with increasing difficulty, enabling progressive robustness learning under controlled distribution shifts. In Stage II, the model is re-aligned with clean task distributions to recover execution fidelity while preserving robustness. We further establish a comprehensive benchmark with 28 perturbation types spanning both textual and visual modalities, grounded in realistic sources of sensor noise, occlusion, and instruction corruption. Extensive experiments on the LIBERO benchmark show that STRONG-VLA consistently improves task success rates across multiple VLA architectures. On OpenVLA, our method achieves gains of up to 12.60% under seen perturbations and 7.77% under unseen perturbations. Notably, similar or larger improvements are observed on OpenVLA-OFT (+14.48% / +13.81%) and pi0 (+16.49% / +5.58%), demonstrating strong cross-architecture generalization. Real-world experiments on an AIRBOT robotic platform further validate its practical effectiveness. These results highlight the importance of decoupled optimization for multimodal robustness and establish STRONG-VLA as a simple yet principled framework for robust embodied control.