PulseAugur
中
实时 21:21:41
English(EN) UEmbed: Unified Sparse and Dense Multimodal Embeddings

新模型统一稀疏和密集多模态嵌入，提升搜索效率

作者 PulseAugur 编辑部 · [4 个来源] ·

研究人员推出 UEmbed，一种新颖的仅解码器多模态嵌入模型，能够在单次因果前向传播中生成稀疏词汇和密集表示。该模型旨在将对现代搜索系统至关重要的稀疏检索与多模态输入统一起来。UEmbed 以 2B、4B 和 9B 的规模发布，其中 9B 版本在 MMEB-v2 和 BEIR 等基准测试中取得了有竞争力的结果，性能优于 RzenEmbed 等现有模型。此外，另一篇独立的研究论文提出了 ReLoop-UME 方法，该方法通过在模型深度上递归地重用参数共享的检索形成块来增强通用多模态嵌入，显著提高了检索速度和有效性。 AI

影响 这些多模态嵌入的进展可以通过提高处理不同数据类型的效率和有效性来增强搜索能力和代理应用。

排序理由 该集群描述了两篇关于多模态嵌入新方法的最新研究论文。

在 Hugging Face Daily Papers 阅读 →

AI 生成摘要 · Google Gemini · 来自 4 个来源。 我们如何撰写摘要 →

新模型统一稀疏和密集多模态嵌入，提升搜索效率

本文如何被排名

Signal score
0 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Research
该集群描述了两篇关于多模态嵌入新方法的最新研究论文。
Source corroboration
4 independent sources
Strong cross-source corroboration — multiple independent publishers covered this within the clustering window.
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
59 days old
Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.

完整方法见我们的编辑标准。

报道来源 [4]

  1. arXiv cs.CL TIER_1 English(EN) · Tingyu Song, Mingxin Li, Yanzhao Zhang, Dingkun Long, Pengjun Xie, Zhijie Nie, Yilun Zhao, Shu Wu ·

    UEmbed：统一的稀疏和密集多模态嵌入

    arXiv:2608.02583v1 Announce Type: cross Abstract: Sparse retrieval underpins modern search systems, from web search to retrieval-augmented generation. Existing work has introduced Learned Sparse Retrieval (LSR) to push beyond exact lexical matching toward richer semantics. Yet LS…

  2. arXiv cs.IR (Information Retrieval) TIER_1 English(EN) · Shu Wu ·

    UEmbed：统一的稀疏和密集多模态嵌入

    Sparse retrieval underpins modern search systems, from web search to retrieval-augmented generation. Existing work has introduced Learned Sparse Retrieval (LSR) to push beyond exact lexical matching toward richer semantics. Yet LSR has so far remained tied to encoder-style bidire…

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

    UEmbed：统一的稀疏和密集多模态嵌入

    Sparse retrieval underpins modern search systems, from web search to retrieval-augmented generation. Existing work has introduced Learned Sparse Retrieval (LSR) to push beyond exact lexical matching toward richer semantics. Yet LSR has so far remained tied to encoder-style bidire…

  4. arXiv cs.CV TIER_1 English(EN) · Shijie Wang, Xiangzhao Hao, Yueti Li, Guangyu Cao, Xinyu Tang, Haiyun Guo ·

    ReLoop-UME：具有可学习检索寄存器的循环深度，用于通用多模态嵌入

    arXiv:2607.28751v1 Announce Type: new Abstract: Universal multimodal embedding (UME) maps heterogeneous multimodal inputs into a shared embedding space. Existing UME models either form embeddings through single forward encoding or add computation through explicit rationale tokens…

相关实体

相关话题