mlsanityData to 14 Oct 2026

VLA Models Are More Generalizable Than You Think: Revisiting Physical and Spatial Modeling

arXiv 2512.02902 ↗First posted 2 Dec 2025cs.RO cs.AI cs.LGcited by 12

On the map: Vision-language-action models → VLA architectures and generalization

Abstract (arXiv)

Vision-language-action (VLA) models achieve strong in-distribution performance but degrade sharply under novel camera viewpoints and visual perturbations. We show that this brittleness primarily arises from misalignment in Spatial Modeling, rather than Physical Modeling. To address this, we propose a one-shot adaptation framework that recalibrates visual representations through lightweight, learnable updates. Our first method, Feature Token Modulation (FTM), applies a global affine transformation to visual tokens and improves Libero viewpoint accuracy from 48.5% to 87.1% with only 4K parameters. Building on this, Feature Linear Adaptation (FLA) introduces low-rank updates to the ViT encoder, achieving 90.8% success with 4.7M parameters -- matching LoRA-scale finetuning at far lower cost. Together, these results reveal substantial untapped robustness in pretrained VLA models and demonstrate that targeted, minimal visual adaptation is sufficient to restore viewpoint generalization.

Key facts. Cited by 12 papers in the map (data through 14 Oct 2026). Placed in VLA architectures and generalization, one of 13 directions in Vision-language-action models. Builds on 22 of its 42 references that are in the map. First posted on arXiv 2 Dec 2025.

Builds on

22 of 42 references.

References · 22

  1. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,380
  2. 2017Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networkscited by 4,897
  3. 2019Parameter-Efficient Transfer Learning for NLPcited by 4,236
  4. 2021Prefix-Tuning: Optimizing Continuous Prompts for Generationcited by 3,915
  5. 2023RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Controlcited by 3,080
  6. 2021Learning to Prompt for Vision-Language Modelscited by 2,281
  7. 2023PaLM-E: An Embodied Multimodal Language Modelcited by 2,071
  8. 2025π0.5: a Vision-Language-Action Model with Open-World Generalizationcited by 1,761
  9. 2022Conditional Prompt Learning for Vision-Language Modelscited by 1,433
  10. 2023LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learningcited by 1,410
  11. 2021BitFit: Simple Parameter-efficient Fine-tuning for Transformer-based Masked Language-modelscited by 1,245
  12. 2019Meta-World: A Benchmark and Evaluation for Multi-Task and Meta Reinforcement Learningcited by 1,204
10 more
  1. 2025Fine-Tuning Vision-Language-Action Models: Optimizing Speed and Successcited by 883
  2. 2023Vision-Language Foundation Models as Effective Robot Imitatorscited by 349
  3. 20243D-VLA: A 3D Vision-Language-Action Generative World Modelcited by 328
  4. 2025LIBERO-Plus: In-depth Robustness Analysis of Vision-Language-Action Modelscited by 303
  5. 2023GELLO: A General, Low-Cost, and Intuitive Teleoperation Framework for Robot Manipulatorscited by 189
  6. 2025LIBERO-PRO: Towards Robust and Fair Evaluation of Vision-Language-Action Models Beyond Memorizationcited by 179
  7. 2024VLABench: A Large-Scale Benchmark for Language-Conditioned Robotics Manipulation with Long-Horizon Reasoning Taskscited by 175
  8. 2025GeoAware-VLA: Implicit Geometry Aware Vision-Language-Action Modelcited by 26
  9. 2025Adapt3R: Adaptive 3D Scene Representation for Domain Transfer in Imitation Learningcited by 17
  10. 2025Physical Autoregressive Model for Robotic Manipulation without Action Pretrainingcited by 12

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 12.

12 papers.

  1. 2026Stable Language Guidance for Vision-Language-Action Modelscited by 14
  2. 2026RADAR: Benchmarking Vision-Language-Action Generalization via Real-World Dynamics, Spatial-Physical Intelligence, and Autonomous Evaluationcited by 4
  3. 2026Escaping the Diversity Trap in Robotic Manipulation via Anchor-Centric Adaptationcited by 4
  4. 2026From Fixed to Free Cameras: Calibration-Free View-Robust Vision-Language-Action Modelcited by 3
  5. 2026TAG: Target-Agnostic Guidance for Stable Object-Centric Inference in Vision-Language-Action Modelscited by 2
  6. 2026Cross-View Action Consistency for Camera-Robust Vision-Language-Action Policiescited by 1
  7. 2026OOWM: Structuring Embodied Reasoning and Planning via Object-Oriented Programmatic World Modelingcited by 0
  8. 2026Domain Arithmetic: One-Shot VLA Adaptation under Environmental Shiftscited by 0
  9. 2026The Moving Eye: Enhancing VLA Spatial Generalization via Hybrid Dynamic Data Collectioncited by 0
  10. 2026Towards Zero-Shot Transfer Across Embodiments For Driving VLAscited by 0
  11. 2026VLA-Scope: Shift-Aware Failure Prediction for Vision-Language-Action Modelscited by 0
  12. 2026Triggering Generalist Reasoning via Predictive Uncertainty for Dual-System VLAcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026GS-VLA: Plug-and-Play Viewpoint Canonicalization for Frozen VLA Policies via Gaussian Splatting
  2. 2025Don't Blind Your VLA: Aligning Visual Representations for OOD Generalization
  3. 2026Not All Layers Need Tuning: Diagnosing and Directing Adaptation in Vision-Language-Action Models
  4. 2026One Token Per Frame: Reconsidering Visual Bandwidth in World Models for VLA Policy
  5. 2025SwiftVLA: Unlocking Spatiotemporal Dynamics for Lightweight VLA Models at Minimal Overhead
  6. 2026StableVLA: Towards Robust Vision-Language-Action Models without Extra Data
  7. 2025MAPS: Preserving Vision-Language Representations via Module-Wise Proximity Scheduling for Better Vision-Language-Action Generalization
  8. 2025VLA-Adapter: An Effective Paradigm for Tiny-Scale Vision-Language-Action Model