analiza aiqbench
This research introduces AVT-Fabric, an adaptive framework that optimizes robotic visuo-tactile reasoning by dynamically selecting tactile evidence based on task difficulty. By leveraging a compact 7B Multimodal Large Language Model, the system achieves higher accuracy and significantly reduced latency compared to larger baselines and passive inference methods. The approach demonstrates that intelligent evidence allocation can enhance efficiency without sacrificing performance in real-world robotic fabric comparison scenarios.
AVT-Fabric: Active Visuo-Tactile Perception via Adaptive Evidence Selection for Efficient Robotic Fabric Comparison
arXiv:2609.21377v1 Announce Type: new Abstract: Robotic fabric comparison needs to actively combine visual appearance and tactile cues. Here, we present AVT-Fabric, an RGB-first framework that allocates tactile evidence according to the difficulty of each comparison. A dual-scale gate evaluates answer-token confidence and raw logit separation to determine whether another force-tagged GelSight observation is needed. Compact textual memory preserves the executed history, and majority voting consolidates the selected predictions. On 400 held-out comparisons, AVT-Fabric achieves 98.0% accuracy with a compact 7B Multimodal Large Language Model (MLLM), surpassing the 94.0% reported by the 90B MLLM-Fabric baseline by 4.0 percentage points while processing only 1.60 of five available stages on average. It improves on matched passive inference by 9.25 percentage points and reduces model-side latency by 61.8%, while also improving on RGB-only accuracy. Four additional MLLM backbones support the generalizability, accuracy, and efficiency of the framework. This framework is also deployed on a real robotic system, achieving 78.1% pairwise ranking accuracy and correct fabric selection in seven of eight application scenarios. AVT-Fabric demonstrates that adaptive evidence selection can improve both the accuracy and efficiency of robotic visuo-tactile reasoning.
Abstract:Robotic fabric comparison needs to actively combine visual appearance and tactile cues. Here, we present AVT-Fabric, an RGB-first framework that allocates tactile evidence according to the difficulty of each comparison. A dual-scale gate evaluates answer-token confidence and raw logit separation to determine whether another force-tagged GelSight observation is needed. Compact textual memory preserves the executed history, and majority voting consolidates the selected predictions. On 400 held-out comparisons, AVT-Fabric achieves 98.0% accuracy with a compact 7B Multimodal Large Language Model (MLLM), surpassing the 94.0% reported by the 90B MLLM-Fabric baseline by 4.0 percentage points while processing only 1.60 of five available stages on average. It improves on matched passive inference by 9.25 percentage points and reduces model-side latency by 61.8%, while also improving on RGB-only accuracy. Four additional MLLM backbones support the generalizability, accuracy, and efficiency of the framework. This framework is also deployed on a real robotic system, achieving 78.1% pairwise ranking accuracy and correct fabric selection in seven of eight application scenarios. AVT-Fabric demonstrates that adaptive evidence selection can improve both the accuracy and efficiency of robotic visuo-tactile reasoning.
Robotics (cs.RO)
arXiv:2609.21377 [cs.RO]
(or arXiv:2609.21377v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2609.21377 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
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.
Text integral sub licența deschisă a publicației · 273 cuvinte
Subiecte conexe
[AINews] not much happened today
Latent Space · 10 sept. 2026
ROMA: LLM System for Real-World Object-Centric Multi-Sensory Active Perception
arXiv · 7 oct. 2026
[AINews] Meta Muse Wonder & Wonder Full: MSL’s Claude 5.5-tier Frontier Model
Latent Space · 13 oct. 2026
NVIDIA GTC Taipei at COMPUTEX: Live Updates on What’s Next in AI
NVIDIA · 21 mai 2026