视觉-语言模型显示出脆弱的路径跟随能力

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-26 04:00

一篇新发表在arXiv上的研究论文强调了当前视觉-语言模型（VLMs）在跟随连续视觉路径方面的显著弱点。研究发现，即使在受控条件下，VLMs也难以维持稳定的路径，尤其是在存在其他附近路径的情况下。这种脆弱性似乎源于视觉编码阶段，在该阶段，路径延续的信息会减弱。通过模型扩展、附加推理或显式跟踪指令来提高性能的尝试收效甚微，这表明现有的VLMs缺乏稳健的路径跟随机制。 AI

影响揭示了VLM路径跟随方面的根本性局限性，可能影响需要视觉导航和场景理解的应用。

排序理由发表在arXiv上的研究论文，详细介绍了当前VLMs的局限性。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 阅读 →

Hyesoo Hong

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 13 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 发表在arXiv上的研究论文，详细介绍了当前VLMs的局限性。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, model release; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Same-day; Cluster formed today. Ranking reflects the current source set at time of score.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.AI TIER_1 English(EN) · Hyesoo Hong, Min Soo Kim, Wonje Jeung, Sangyeon Yoon, Dongjae Jeon, Albert No · 2026-10-26 04:00

VLMs 无痕追踪：诊断视觉路径跟随中的失败

arXiv:2605.15672v2 Announce Type: replace-cross Abstract: Vision-language models (VLMs) perform strongly on multimodal benchmarks, yet their ability to reliably follow continuous visual structure remains unclear. We focus on path following, a basic operation that requires followi…

本文如何被排名

报道来源 [1]

VLMs 无痕追踪：诊断视觉路径跟随中的失败

相关话题