aiqbench analysis
FabriVLA introduces a lightweight Vision-Language-Action model optimized for robotic manipulation, achieving state-of-the-art performance with significantly fewer parameters than standard architectures. The accompanying Joint Conformal Action Chunk Calibration method provides rigorous uncertainty quantification, enabling risk-aware execution and safer multi-task deployment without heavy computational overhead. This work addresses critical bottlenecks in practical robotics by balancing efficiency with auditable confidence levels for action planning.
FabriVLA: A Lightweight Vision-Language-Action Model with Conformal Action Chunk Uncertainty
arXiv:2607.08575v3 Announce Type: replace Abstract: Vision-Language-Action (VLA) models have become a leading paradigm for general purpose robotic manipulation, but their computational cost and limited uncertainty awareness hinder practical deployment. We present FabriVLA, a lightweight VLA that fuses shallow and intermediate VLM layers to preserve fine-grained visual features, and gates self-attention among action tokens so that its flow matching head admits inter step structure only as far as training warrants. Trained end-to-end in a single stage, FabriVLA reaches a state-of-the-art 90.0\% average success on Meta-World MT50 with only 0.88B parameters. We further introduce Joint Conformal Action Chunk Calibration (JCAC), a post-training method that augments a frozen policy with a lightweight residual scale head. From a single policy query, JCAC turns a learned elementwise error scale into a set that covers the whole executed action prefix at a user chosen confidence level, 3.3$\times$ tighter in mean radius than an unconditional conformal set. On LIBERO-Safety, these bounds rank rollouts by risk before execution, supporting risk ranked review. Together, FabriVLA and JCAC provide a lightweight and auditable framework for multi-task manipulation with calibrated action uncertainty.
Abstract:Vision-Language-Action (VLA) models have become a leading paradigm for general purpose robotic manipulation, but their computational cost and limited uncertainty awareness hinder practical deployment. We present FabriVLA, a lightweight VLA that fuses shallow and intermediate VLM layers to preserve fine-grained visual features, and gates self-attention among action tokens so that its flow matching head admits inter step structure only as far as training warrants. Trained end-to-end in a single stage, FabriVLA reaches a state-of-the-art 90.0\% average success on Meta-World MT50 with only 0.88B parameters. We further introduce Joint Conformal Action Chunk Calibration (JCAC), a post-training method that augments a frozen policy with a lightweight residual scale head. From a single policy query, JCAC turns a learned elementwise error scale into a set that covers the whole executed action prefix at a user chosen confidence level, 3.3$\times$ tighter in mean radius than an unconditional conformal set. On LIBERO-Safety, these bounds rank rollouts by risk before execution, supporting risk ranked review. Together, FabriVLA and JCAC provide a lightweight and auditable framework for multi-task manipulation with calibrated action uncertainty.
Robotics (cs.RO)
arXiv:2607.08575 [cs.RO]
(or arXiv:2607.08575v3 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2607.08575 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 267 words
Related coverage
Adaptive Conformal Redistribution for Inter-class Transitional Uncertainty in Medical Image Classification
arXiv · Sep 16, 2026
TurboVLA: Real-Time Vision-Language-Action Model at 32 Hz on an RTX 4090 with <1 GB VRAM
arXiv · Jul 30, 2026
RT-VLA: Real-Time Vision-Language-Action Models via Knowledge Distillation
arXiv · Jun 16, 2026
Learning What to Say to Your VLA: Mostly Harmless Vision Language Action Model Steering
arXiv · Jun 12, 2026