PulseAugur
实时 11:24:31
English(EN) Does Forgetting Transfer Across Modalities? A Real-World Benchmark for Cross-Modal Knowledge Unlearning Evaluation

新的基准 UNLINK-VL 评估了 VLMs 中的跨模态知识遗忘

研究人员推出了 UNLINK-VL,这是一个旨在评估知识从视觉语言模型 (VLMs) 中有效移除程度的新基准。该基准侧重于遗忘在不同模态之间的迁移,填补了当前研究中通常只考虑单模态遗忘的空白。实验表明,虽然多模态遗忘对于文本评估是有效的,但仅文本遗忘在视觉或跨模态测试时表现不佳。这凸显了如果只进行模态内评估,可能会高估遗忘的有效性。 AI

影响 强调了在 AI 遗忘中进行稳健的跨模态评估的必要性,以避免高估其有效性。

排序理由 该集群描述了一篇介绍 AI 模型遗忘评估基准的新学术论文。

在 arXiv cs.AI 阅读 →

AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →

新的基准 UNLINK-VL 评估了 VLMs 中的跨模态知识遗忘

报道来源 [2]

  1. arXiv cs.AI TIER_1 English(EN) · Chunlin Liu, Junnian Chen, Haitong Jiang, Jianyu Zhao, Yingsen Pang, Jingchen Li, Jiabiao He, Youming Lu, Jinhe Bi, Yuntao Du ·

    遗忘是否能跨模态迁移?用于跨模态知识遗忘评估的真实世界基准

    arXiv:2608.03791v1 Announce Type: new Abstract: Vision-Language Models (VLMs), like Large Language Models (LLMs), may memorize sensitive, copyrighted, or harmful knowledge from their pretraining corpora. Removing such knowledge is essential for building trustworthy AI systems. Ho…

  2. Hugging Face Daily Papers TIER_1 English(EN) ·

    Does Forgetting Transfer Across Modalities? A Real-World Benchmark for Cross-Modal Knowledge Unlearning Evaluation

    Vision-Language Models (VLMs), like Large Language Models (LLMs), may memorize sensitive, copyrighted, or harmful knowledge from their pretraining corpora. Removing such knowledge is essential for building trustworthy AI systems. However, existing studies primarily focus on forge…