mlsanityData to 14 Oct 2026

Neural Introspection Gating for Adaptive KV-Cache Reuse in Vision-Language-Action Models

arXiv 2608.10824 ↗First posted 11 Aug 2026cs.RO cs.CVcited by 1

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

also in 1 more
Abstract (arXiv)

Vision-Language-Action(VLA) models map camera images and language instructions directly to motor commands through a single autoregressive transformer. In real-time control, they still spend substantial compute recomputing key-value(KV) representations for visual tokens that barely change across neighboring frames. Recent work such as VLA-Cache reduces that cost by reusing KV states for visually static patches, but its policy relies only on observation-space heuristics and does not account for the model's own uncertainty. We propose Gated VLA-Cache, a lightweight, training-free extension that augments visual-similarity caching with neural introspection. The method monitors the logit margin between the top two predicted action tokens, a zero-cost confidence signal available during decoding. When the margin drops below a threshold, the cache is invalidated and a full recompute is triggered. Evaluated on four LIBERO benchmark suites with both OpenVLA and OpenVLA-OFT, Gated VLA-Cache improves reliability when blind caching hurts. On LIBERO-Goal and LIBERO-Long, it recovers over 100% of the lost accuracy while retaining 80% of the compute savings.

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 22 of its 24 references that are in the map. First posted on arXiv 11 Aug 2026.

Builds on

22 of 24 references.

Closest ancestors · 1

Same direction, newest first

  1. 2025VLA-Cache: Efficient Vision-Language-Action Manipulation via Adaptive Token Caching4 Feb

Other references · 21

  1. 2017Attention Is All You Needcited by 60,361
  2. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 13,028
  3. 2022Flow Matching for Generative Modelingcited by 5,702
  4. 2023Mamba: Linear-Time Sequence Modeling with Selective State Spacescited by 4,718
  5. 2015Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learningcited by 4,448
  6. 2016Simple and Scalable Predictive Uncertainty Estimation using Deep Ensemblescited by 3,926
  7. 2023RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Controlcited by 3,080
  8. 2024OpenVLA: An Open-Source Vision-Language-Action Modelcited by 2,887
  9. 2023Diffusion Policy: Visuomotor Policy Learning via Action Diffusioncited by 2,838
  10. 2024π0: A Vision-Language-Action Flow Model for General Robot Controlcited by 2,407
  11. 2016A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networkscited by 2,385
  12. 2023Efficient Streaming Language Models with Attention Sinkscited by 2,186
9 more
  1. 2023LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learningcited by 1,410
  2. 2025Fine-Tuning Vision-Language-Action Models: Optimizing Speed and Successcited by 883
  3. 2023Open X-Embodiment: Robotic Learning Datasets and RT-X Modelscited by 862
  4. 2024An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Modelscited by 642
  5. 2025SmolVLA: A Vision-Language-Action Model for Affordable and Efficient Roboticscited by 530
  6. 2024Prismatic VLMs: Investigating the Design Space of Visually-Conditioned Language Modelscited by 339
  7. 2024TinyVLA: Towards Fast, Data-Efficient Vision-Language-Action Models for Robotic Manipulationcited by 339
  8. 2024PyramidDrop: Accelerating Your Large Vision-Language Models via Pyramid Visual Redundancy Reductioncited by 294
  9. 2024RoboMamba: Efficient Vision-Language-Action Model for Robotic Reasoning and Manipulationcited by 103

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. 2025KV-Efficient VLA: A Method to Speed up Vision Language Models with RNN-Gated Chunked KV Cache
  2. 2026Text-Vision Synergistic Token Caching: A Training-Free Framework for Efficient Vision-Language-Action Inference
  3. 2025Think Twice, Act Once: Token-Aware Compression and Action Reuse for Efficient Inference in Vision-Language-Action Models
  4. 2025EfficientVLA: Training-Free Acceleration and Compression for Vision-Language-Action Models
  5. 2026ActionCache: Training-Free Acceleration for Vision-Language-Action Models with Action Caching and Refinement
  6. 2026Not All Layers Need Tuning: Diagnosing and Directing Adaptation in Vision-Language-Action Models
  7. 2026DepthCache: Depth-Guided Training-Free Visual Token Merging for Vision-Language-Action Model Inference
  8. 2025MAPS: Preserving Vision-Language Representations via Module-Wise Proximity Scheduling for Better Vision-Language-Action Generalization