BMTL 提出 Banach 空间域适应方法：用 L1 密度正则对齐特征流形
The comparison of density estimation for transfer learning in Banach, Hilbert, Euclidean spaces
做域适应的人可以看看这篇：把 L2 换成 L1 Banach 范数正则，靠 Scheffe 定理直接收紧 TV 距离，在 Office+Caltech10 的 12 个迁移上比 DMDA 平均高 3.68%。论文还把三种几何摆在一起对比，理论味比较足。
arXiv 新发布的预印本总结了论文 Non-Hilbertian Domain Adaptation，提出 Banach Manifold Transfer Learning（BMTL），针对现有域适应方法假定 Hilbert 几何、用对称 L2 散度衡量分布差异的局限做出改进。BMTL 包含两个耦合设计：特征嵌入在测地流形而非平坦欧氏空间上；密度匹配正则项采用 L1 Banach 范数，依据 Scheffe 定理，其最小化等价于收缩域密度间的 Total Variation（TV）距离。
技术机理上，TV 距离对样本空间各区域均匀加权且有界，而 L2 形式对偏差平方会放大尾部误差，过度强调不能反映真实分布偏移的密度离群点。在 MNIST-USPS、COIL20、Office+Caltech10 三个基准上，L1 变体在每个任务上都超过同条件下的 L2 消融版本，例如 USPS->MNIST 达到 92.44% 对 89.80%。
与最强基线 DMDA 相比，BMTL 在 12 个 Office+Caltech10 迁移任务上平均高出 3.68%。论文还给出了欧氏、Hilbert、Banach 三种几何的显式对比，为域适应中的几何选择提供了理论框架，对在目标标签稀缺场景做迁移学习的研究者有直接参考价值。
The comparison of density estimation for transfer learning in Banach, Hilbert, Euclidean spaces
Domain adaptation or DA methods are indispensable when target labels or computational resources are scarce, yet the domain kernel density and subspace alignment approaches completely assume a Hilbertian geometry. They measure distributional discrepancy with symmetric L2 divergences over Euclidean feature spaces or reproducing kernel Hilbert spaces (RKHS). This preprint summarizes the manuscript Non-Hilbertian Domain Adaptation. Total Variation Alignment via L1 Density Regularization on Feature Manifolds, which proposes Banach Manifold Transfer Learning or BMTL. BMTL makes two coupled design decisions. The first one is that features are embedded on a geodesic manifold rather than treated as points of a flat Euclidean space, and the second one is that the density matching regularizer is formulated in an L1 Banach norm, whose minimization is equivalent, by Scheffe's theorem, to contracting the Total Variation or TV distance between domain densities, a bounded metric that weights all regions of the sample space uniformly. The L2 formulation, by contrast, squares deviations and therefore amplifies errors in the tails, over emphasizing density outliers that do not reflect meaningful distributional shift. Experiments on MNIST-USPS, COIL20, and Office+Caltech10 show consistent gains; the L1 variant beats an otherwise identical L2 ablation on every task, for example 92.44% vs. 89.80% on USPS->MNIST and surpasses the strongest baseline DMDA by +3.68% on average over twelve Office+Caltech transfers. This document situates the contribution with an explicit comparison of Euclidean, Hilbert, and Banach geometries.
本条情报归属 AITOP「论文」核心追踪栏目。在跨区域产业观察坐标中，该动态直接关联「AI 开发者工具链、Coding Agent 与工程化实践」研究主线，反映出当前技术演进与商业落地的最新风向。涉及核心实体与技术议题包括：域适应、迁移学习、Total Variation、Banach空间，读者可通过下方标签进入对应专题页查阅历史脉络与关联对比。
信源核验说明：本文原始线索采编自「arXiv cs.AI」（@Maryam Alsadat Kazemian, Azam Bastanfard），经 AITOP 多维评分管线结构化提炼。