UNIEGO：它和通用视觉自监督的关系在于：用 proxy 中介层融合 ego/exo、RGB/depth/skeleton 与多个 foundation teacher，适合作为视频 SSL 蒸馏读物
中高相关；详见方法、贡献和实验边界。
UNIEGO: Proxies as Mediators for Unified Egocentric Video Representation Learning arXiv 新增 + video representation + multi-teacher distillation 原文链接
编号2606.20559优先级P1类别arXiv 新增 + video representation + multi-teacher distillation会议arXiv 新增 + video representation + multi-teacher distillation方法用 proxy 中介层融合 ego/exo、RGB/depth/skeleton 与多个 foundation teacher，适合作为视频 SSL 蒸馏读物来源arXiv / OpenReview
先说结论。它和通用视觉自监督的关系在于：用 proxy 中介层融合 ego/exo、RGB/depth/skeleton 与多个 foundation teacher，适合作为视频 SSL 蒸馏读物。 中高相关；详见方法、贡献和实验边界。
Level-I: Proxy Learning Level-II: Proxy Merging and Selective Proxy Distillation Figure 2:Level-I: Proxy Learning Level-II: Proxy Merging and Selective Proxy Distillation Figure 2: Overview of UNIEGO. UNIEGO learns a unified egocentric encoder through a twolevel proxy-mediated distillation framework. In Level-I (left), heterogeneous teachers spanning viewpoints, modalities, and foundation representations independently supervise egocentric proxy models, converting diverse teacher signals into a homogeneous proxy space. In Level-II (right), the proxy parameters are first merged to initialize the unified model, after which Selective Proxy Distillation (SPD) performs sample-wise reliability filtering and distills only from proxies that are correct and confident. At inference, only the final UNIEGO model operates using egocentric input.这张图概括 UNIEGO 的整体方法流程。阅读时先看模块之间传递的训练信号，再看作者如何把目标拆成可优化的子问题。Figureure 1 · : (a) Naive multi-teacher distillation with heterogeneous teachers forFigure 1: (a) Naive multi-teacher distillation with heterogeneous teachers for learning unified egocentric representations results in representational gaps and conflicting gradients, as illustrated in (b). (c) In contrast, our proposed UNIEGO adopts a hierarchical distillation framework that mitigates these limitations through proxy-mediated learning, as shown in (d). Black dashed arrows illustrate the effects of this framework, shifting teachers into a unified representation space.这张图概括 UNIEGO 的整体方法流程。阅读时先看模块之间传递的训练信号，再看作者如何把目标拆成可优化的子问题。
核心问题
它和通用视觉自监督的关系在于：用 proxy 中介层融合 ego/exo、RGB/depth/skeleton 与多个 foundation teacher，适合作为视频 SSL 蒸馏读物。
方法拆解
用 proxy 中介层融合 ego/exo、RGB/depth/skeleton 与多个 foundation teacher，适合作为视频 SSL 蒸馏读物
