From Fragmentation to Fusion: A Survey and Roadmap for Omnimodal Embeddings

Computer Science > Computation and Language arXiv:2611.00172 (cs) [Submitted on 30 Oct 2026] Title: From Fragmentation to Fusion: A Survey and Roadmap for Omnimodal Embeddings.

计算机科学 > 计算与语言 arXiv:2611.00172 (cs) [提交于 2026 年 10 月 30 日] 标题：从碎片化到融合：全模态嵌入的综述与路线图。

Abstract: Embedding models underpin virtually every modern AI system, from search engines to recommendation systems. Despite shared mathematical foundations, research on embedding models remains fragmented by modality: the communities develop largely in isolation, seldom exchanging insights.

摘要：嵌入模型支撑着几乎所有现代人工智能系统，从搜索引擎到推荐系统。尽管拥有共同的数学基础，但嵌入模型的研究仍按模态呈现碎片化状态：各研究社区大多处于孤立发展中，鲜有交流见解。

We argue that this fragmentation obscures a deeper convergence. Driven by contrastive learning, Transformer architectures, and a granularity shift, embedding research across modalities is converging toward a common set of principles, even as modality-specific challenges remain underexplored.

我们认为，这种碎片化掩盖了更深层次的融合。在对比学习、Transformer 架构和粒度转变的驱动下，跨模态的嵌入研究正趋向于一套共同的原则，尽管针对特定模态的挑战仍有待深入探索。

To reason across this convergence without conflating it, we propose a framework of three progressively harder capabilities: coverage, comparability, and compositionality. Three paths toward unification satisfy these in turn: transformation broadens coverage, alignment enables comparability, and joint embedding, the terminus we advocate, aims at compositionality.

为了在不混淆的前提下对这种融合进行推理，我们提出了一个包含三个难度递增能力的框架：覆盖性、可比性和组合性。通向统一的三条路径依次满足这些要求：转换扩大了覆盖范围，对齐实现了可比性，而我们所倡导的终极目标——联合嵌入，则旨在实现组合性。

This position paper traces the resulting evolution, providing the first unified, cross-cutting treatment of omnimodal embedding research.

这篇立场论文追溯了由此产生的演变过程，为全模态嵌入研究提供了首个统一的、跨领域的论述。