MOVE：多模态开放世界图学习验证与扩展框架

1. 12:00
   arXiv · Machine Learning Theory

   MOVE：面向图学习的多模态开放世界类别验证与扩展框架

   MOVE 提出多模态开放世界类别验证与扩展框架，通过联合视觉 token、文本属性与图上下文识别无法归入现有类别的节点，并利用多模态 LLM 生成候选类别，仅在多模态证据一致支持且不引入冗余类别时才扩展类别空间。实验表明 MOVE 在未知节点识别、开放域标注和下游图学习任务上平均提升 11.87%。

Current heat 9·Comparable peak 10（Oct 2）·Comparable change over 24 hours –

The trend compares only the same participants observed continuously; its range may be smaller than the current heat count. Move or click on the chart to inspect hourly heat; use the left and right arrow keys to switch.