发表机构
Florida Atlantic University（佛罗里达大西洋大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究弱监督乳腺钼靶成像中特征退化问题，引入基于梯度的正交潜在分解用于分层变分自编码器，划分潜在空间，通过实验得出模型相关指标，经潜在消融等验证，揭示高维空间中粗监督信号对细粒度特征的影响。
AI 中文摘要
弱监督分层模型存在持续的不对称性：粗病变类型特征在重建时得以保留，而细粒度恶性线索会退化，这对乳腺癌筛查管道的临床可靠性有直接影响。我们引入基于梯度的正交潜在分解用于分层变分自编码器（H-VAEs），以机械地解释这种不对称性。潜在空间被划分为由粗监督梯度塑造的任务对齐组件（$z_1$）和捕获剩余表示能力的正交残差（$z_{\text{res}}$）。在来自CBIS-DDSM的3550个乳腺钼靶感兴趣区域（ROIs）上，只有约4.4%的潜在幅度与监督梯度对齐，约95.6%在正交残差中，细粒度病理预测主要依赖于此。该模型实现了第一阶段AUC为0.866，第二阶段AUC为0.552，重建稳定性差距为$\Delta_{\text{diag}} = 5\%$（$p = 0.005$），分类差距为$\Delta_{\text{AUC}} = 0.314$（$p < 0.001$）。潜在消融证实两个任务的特征都大量存在于$z_{\text{res}}$中，从结构上解释了为什么重建会不成比例地降低病理稳定性。与多实例学习（MIL）和多任务学习（MTL）的比较证实了跨架构和模态的泛化。这些发现表明，在高维空间中，单个粗监督信号仅隔离出一个稀疏的一维潜在方向，迫使关键的细粒度特征进入易受影响的残差子空间。
英文摘要
Weakly supervised hierarchical models exhibit a persistent asymmetry: coarse lesion-type features are preserved under reconstruction while fine-grained malignancy cues degrade---a pattern with direct consequences for the clinical reliability of breast cancer screening pipelines. We introduce gradient-based orthogonal latent decomposition for hierarchical Variational Autoencoders~(H-VAEs) to mechanistically explain this asymmetry. The latent space is partitioned into a task-aligned component~($z_1$), shaped by coarse supervisory gradients, and an orthogonal residual~($z_{\text{res}}$) capturing remaining representational capacity. On~3,550 mammographic Regions of Interest~(ROIs) from CBIS-DDSM, only~$\sim$4.4\% of latent magnitude aligns with supervisory gradients, leaving~$\sim$95.6\% in the orthogonal residual upon which fine-grained pathology prediction primarily depends. The model achieves Stage-1~AUC~0.866 and Stage 2~AUC~0.552, with a reconstruction stability gap of $Δ_{\text{diag}}=5\%$ ($p=0.005$) and a classification gap of $Δ_{\text{AUC}}=0.314$ ($p{<}0.001$). Latent ablation confirms that features for both tasks reside heavily in~$z_{\text{res}}$, structurally explaining why reconstruction degrades pathology stability disproportionately. Comparisons with Multi-Instance Learning~(MIL) and Multi-Task Learning~(MTL) confirm generalization across architectures and modalities. These findings reveal that in high-dimensional spaces, a single coarse supervisory signal isolates only a sparse 1D latent direction, forcing critical fine-grained features into the vulnerable residual subspace.
