arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.15873cs.ROcs.AI
收藏

测试时适应用于触觉-视觉-语言模型

Test-Time Adaptation for Tactile-Vision-Language Models

Shenzhen Technology University(深圳技术大学);New York University(纽约大学);Shenzhen University(深圳大学);Tsinghua University(清华大学)

Chuyang Ye, Haoxian Jing, Qinting Jiang, Yixi Lin, Qiang Li, Xing Tang, Jingyan Jiang

AI总结:

本文提出了一种可靠性感知框架,用于提升触觉-视觉-语言模型在测试时的适应能力,通过估计各模态可靠性来过滤不可靠样本并优化融合过程,有效提升了在模态损坏情况下的准确率。

摘要:

触觉-视觉-语言(TVL)模型越来越多地部署在现实世界的机器人和多模态感知任务中,其中测试时的分布偏移是不可避免的。现有的测试时适应(TTA)方法在单模态设置中提供过滤功能,但缺乏对异步跨模态偏移下各模态可靠性显式处理的能力,这使它们在某些模态变得不可靠时变得脆弱。我们研究了在这样的偏移下TVL模型的TTA,并提出了一种可靠性意识框架,该框架从预测不确定性及扰动响应中估计各模态的可靠性。此共享的可靠性信号用于(i)过滤不可靠的测试样本,(ii)自适应融合触觉、视觉和语言特征,(iii)用可靠性引导的目标正则化测试时优化。在TAG-C基准和其他额外的TVL场景上,我们的方法始终优于强大的TTA基线,实现了在严重模态损坏情况下准确率提升高达49.9%,突显了显式模态级可靠性建模对于稳健测试时适应的重要性。

英文摘要:

Tactile-vision-language (TVL) models are increasingly deployed in real-world robotic and multimodal perception tasks, where test-time distribution shifts are unavoidable. Existing test-time adaptation (TTA) methods provide filtering in unimodal settings but lack explicit treatment of modality-wise reliability under asynchronous cross-modal shifts, leaving them brittle when some modalities become unreliable. We study TTA for TVL models under such shifts and propose a reliability-aware framework that estimates per-modality reliability from prediction uncertainty and perturbation-based responses. This shared reliability signal is used to (i) filter unreliable test samples, (ii) adaptively fuse tactile, visual, and language features, and (iii) regularize test-time optimization with a reliability-guided objective. On the TAG-C benchmark and additional TVL scenarios, our approach consistently outperforms strong TTA baselines, achieving accuracy gains of up to 49.9\% under severe modality corruptions, underscoring the importance of explicit modality-wise reliability modeling for robust test-time adaptation.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑