Omni-modal retrieval promises a single embedding space for text, image, video, document, and audio inputs, but building such a unified retriever is difficult since these modalities differ in data distribution, architecture, and optimization dynamics. In this work, we present Conan-embedding-v3, a decouple--fuse--recover framework for omni-modal retrieval. Conan-embedding-v3 first trains modality specialists independently and fuses their task vectors into a single dense backbone, a strategy we call Decoupled Specialist Fusion. We show that this fusion composes visual, video, and document retrieval capabilities, but also exposes a failure mode for projector-based modalities: when audio is attached through an external encoder and projector, fusing the backbone leaves the projector calibrated to the audio-specialist backbone, causing a large audio retrieval regression despite copying all audio-specific modules unchanged. We call this failure Projector Drift. To repair it, Conan-embedding-v3 applies Projector Recovery (i.e., full-parameter fine-tuning of the projector while keeping the backbone frozen) followed by balanced multi-modal rehearsal. The resulting model supports these retrieval pathways in one backbone, achieving 74.9 scores on MMEB while obtaining 55.61 on the 30-task MAEB audio suite.

翻译：全模态检索旨在为文本、图像、视频、文档和音频输入建立统一的嵌入空间，但由于这些模态在数据分布、架构和优化动力学上存在差异，构建此类统一检索器颇具挑战。本文提出Conan-embedding-v3——一种面向全模态检索的解耦-融合-恢复框架。该框架首先独立训练各模态专家，然后通过"解耦专家融合"策略将其任务向量整合至单一稠密骨干网络。实验表明，该融合方式能组合视觉、视频与文档检索能力，但暴露出基于投影仪的模态的失效模式：当通过外部编码器与投影仪接入音频时，融合骨干网络会使投影仪仍校准于音频专有骨干网络，导致即便完整保留所有音频专用模块，音频检索性能仍出现大幅下降（即"投影仪偏移"失效）。为解决此问题，Conan-embedding-v3采用投影仪恢复技术（冻结骨干网络同时全参数微调投影仪）并辅以平衡多模态重放。最终模型在单一骨干网络支持下实现多路径检索，分别在MMEB基准上取得74.9分，在包含30个任务的MAEB音频套件中取得55.61分。