We introduce H-SPLID, a novel algorithm for learning salient feature representations through the explicit decomposition of salient and non-salient features into separate spaces. We show that H-SPLID promotes learning low-dimensional, task-relevant features. We prove that the expected prediction deviation under input perturbations is upper-bounded by the dimension of the salient subspace and the Hilbert-Schmidt Independence Criterion (HSIC) between inputs and representations. This establishes a link between robustness and latent representation compression in terms of the dimensionality and information preserved. Empirical evaluations on image classification tasks show that models trained with H-SPLID primarily rely on salient input components, as indicated by reduced sensitivity to perturbations affecting non-salient features, such as image backgrounds. Our code is available at https://github.com/neu-spiral/H-SPLID.

翻译：本文提出H-SPLID，一种通过将显著特征与非显著特征显式分解至独立空间来学习显著特征表示的新算法。我们证明H-SPLID能够促进低维任务相关特征的学习。从理论上，我们证明了输入扰动下的预期预测偏差受显著子空间维度与输入-表示间希尔伯特-施密特独立性准则（HSIC）的上界约束，从而在维度与信息保留层面建立了鲁棒性与隐表示压缩的关联。在图像分类任务上的实证评估表明，采用H-SPLID训练的模型主要依赖显著输入成分，其具体表现为对影响非显著特征（如图像背景）的扰动敏感性降低。代码已开源：https://github.com/neu-spiral/H-SPLID。