mlsanityData to 14 Oct 2026

Reducing Temporal Redundancy for Efficient Vision-Language-Action Inference

arXiv 2607.12287 ↗First posted 14 Jul 2026cs.ROcited by 1

On the map: Vision-language-action models → Efficient VLA inference and adaptation

Abstract (arXiv)

Vision-Language-Action (VLA) models exhibit strong generalization for robotic manipulation, yet their high inference latency limits real time deployment. We identify two primary sources of temporal redundancy in existing VLA pipelines: repeated visual encoding of highly similar consecutive frames and multi step iterative sampling in diffusion based policies. To address this, we propose a system level acceleration strategy that reduces computation in both perception and action generation. On the perception side, we incrementally update only tokens corresponding to dynamic scene regions instead of re-encoding entire frames. On the policy side, we compress diffusion sampling into a compact 2-step schedule through efficiency oriented training while preserving action precision. Experiments on Libero, RobotWin, and Real Robot Platforms demonstrate over 2 times speedup while maintaining high performance, achieving up to 98% success rate on general manipulation benchmarks. Our codes will be released on Github.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Placed in Efficient VLA inference and adaptation, one of 13 directions in Vision-language-action models. Builds on 36 of its 42 references that are in the map. First posted on arXiv 14 Jul 2026.

Builds on

36 of 42 references.

Closest ancestors · 4

Same direction, newest first

  1. 2025VLASH: Real-Time VLAs via Future-State-Aware Asynchronous Inference30 Nov
  2. 2025EfficientVLA: Training-Free Acceleration and Compression for Vision-Language-Action Models11 Jun
  3. 2025Think Twice, Act Once: Token-Aware Compression and Action Reuse for Efficient Inference in Vision-Language-Action Models27 May
  4. 2025FAST: Efficient Action Tokenization for Vision-Language-Action Models16 Jan

Other references · 32

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  2. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  3. 2021Swin Transformer: Hierarchical Vision Transformer using Shifted Windowscited by 12,139
  4. 2022Flow Matching for Generative Modelingcited by 5,702
  5. 2022BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generationcited by 4,322
  6. 2017MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applicationscited by 3,845
  7. 2023Sigmoid Loss for Language Image Pre-Trainingcited by 3,109
  8. 2023RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Controlcited by 3,080
  9. 2024OpenVLA: An Open-Source Vision-Language-Action Modelcited by 2,887
  10. 2023Diffusion Policy: Visuomotor Policy Learning via Action Diffusioncited by 2,838
  11. 2024π0: A Vision-Language-Action Flow Model for General Robot Controlcited by 2,407
  12. 2025π0.5: a Vision-Language-Action Model with Open-World Generalizationcited by 1,761
20 more
  1. 2023Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardwarecited by 1,568
  2. 2023LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learningcited by 1,410
  3. 2017ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devicescited by 1,200
  4. 2022Planning with Diffusion for Flexible Behavior Synthesiscited by 1,061
  5. 2022Token Merging: Your ViT But Fastercited by 1,021
  6. 2022Diffusion Models in Vision: A Surveycited by 1,015
  7. 2022DPM-Solver++: Fast Solver for Guided Sampling of Diffusion Probabilistic Modelscited by 938
  8. 2024SnapKV: LLM Knows What You are Looking for Before Generationcited by 880
  9. 2020A Survey on Visual Transformercited by 654
  10. 2024TOFU: A Task of Fictitious Unlearning for LLMscited by 533
  11. 2025RoboTwin 2.0: A Scalable Data Generator and Benchmark with Strong Domain Randomization for Robust Bimanual Robotic Manipulationcited by 518
  12. 2024Mobile ALOHA: Learning Bimanual Mobile Manipulation with Low-Cost Whole-Body Teleoperationcited by 391
  13. 2024SparseVLM: Visual Token Sparsification for Efficient Vision-Language Model Inferencecited by 391
  14. 2024TinyVLA: Towards Fast, Data-Efficient Vision-Language-Action Models for Robotic Manipulationcited by 339
  15. 2025X-VLA: Soft-Prompted Transformer as Scalable Cross-Embodiment Vision-Language-Action Modelcited by 303
  16. 2024A Survey on Vision-Language-Action Models for Embodied AIcited by 279
  17. 2022Multimodal Token Fusion for Vision Transformerscited by 136
  18. 2024DeeR-VLA: Dynamic Inference of Multimodal Large Language Models for Efficient Robot Executioncited by 117
  19. 2025Variation-aware Vision Token Dropping for Faster Large Vision-Language Modelscited by 27
  20. 2025DART: Differentiable Dynamic Adaptive Region Tokenizer for Vision Foundation Modelscited by 4

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026WAM-Cache: Staleness-Bounded KV Reuse for Efficient World Action Modelscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Learning to Accelerate Vision-Language-Action Models through Adaptive Visual Token Caching
  2. 2026AC^2-VLA: Action-Context-Aware Adaptive Computation in Vision-Language-Action Models for Efficient Robotic Manipulation
  3. 2025VLA-Cache: Efficient Vision-Language-Action Manipulation via Adaptive Token Caching
  4. 2026A1: A Fully Transparent Open-Source, Adaptive and Efficient Truncated Vision-Language-Action Model
  5. 2025FASTer: Toward Efficient Autoregressive Vision Language Action Modeling via Neural Action Tokenization
  6. 2026Finetuning Vision-Language-Action Models Requires Fewer Layers Than You Think
  7. 2025Action-aware Dynamic Pruning for Efficient Vision-Language-Action Manipulation
  8. 2026ElegantVLA: Learning When to Think for Efficient Vision-Language-Action Models