提出可解释的隐私对齐框架，动态识别视觉模型中敏感特征区域。
Bodhi VLM: Privacy-Alignment Modeling for Hierarchical Visual Representations in Vision Backbones and VLM Encoders via Bottom-Up and Top-Down Feature Search
- 通过上下文聚类定位敏感特征，结合自底向上与自顶向下搜索策略。
- 在多个检测器和视觉语言模型上验证，输出与隐私预算匹配的可解释信号。
- 适合关注模型隐私保护与可解释性的研究人员使用。
为保障学习系统中的隐私，通常需向分层视觉表示注入噪声；核心挑战在于如何建模此类扰动与预设隐私预算之间的对齐关系，且具备跨视觉骨干网络和视觉-语言模型（VLMs）的适用性。本文提出Bodhi VLM，一种面向分层神经表示的隐私对齐建模框架：(1) 基于NCP与MDAV聚类方法，将敏感概念与层级特征关联；(2) 采用自底向上（BUA）与自顶向下（TDA）策略，在多尺度表示（如特征金字塔或视觉编码器层）中定位敏感特征区域；(3) 设计期望-最大化隐私评估（EMPA）模块，通过比较拟合的敏感特征分布与用户指定参考分布（如尺度为 $c/ε$ 的拉普拉斯或高斯分布），生成可解释的预算对齐信号。该信号为相对参考值，非正式微分隐私估计量。我们形式化了在分层结构上的BUA/TDA，并在目标检测器（YOLO、PPDPTS、DETR）及视觉语言模型的视觉编码器（CLIP、LLaVA、BLIP）上验证框架有效性。结果显示BUA与TDA趋势一致，EMPA在设定条件下提供稳定对齐信号。与通用差异基线（卡方、KL、MMD）及任务相关基线（MomentReg、NoiseMLE、Wass-1）对比，结果以多种子均值±标准差报告，置信区间见补充材料。本工作贡献了一个可学习、可解释的隐私对齐建模视角，而非仅事后审计。
原文摘要 · Abstract (English)
Learning systems that preserve privacy often inject noise into hierarchical visual representations; a central challenge is to \emph{model} how such perturbations align with a declared privacy budget in a way that is interpretable and applicable across vision backbones and vision--language models (VLMs). We propose \emph{Bodhi VLM}, a \emph{privacy-alignment modeling} framework for \emph{hierarchical neural representations}: it (1) links sensitive concepts to layer-wise grouping via NCP and MDAV-based clustering; (2) locates sensitive feature regions using bottom-up (BUA) and top-down (TDA) strategies over multi-scale representations (e.g., feature pyramids or vision-encoder layers); and (3) uses an Expectation-Maximization Privacy Assessment (EMPA) module to produce an interpretable \emph{budget-alignment signal} by comparing the fitted sensitive-feature distribution to an evaluator-specified reference (e.g., Laplace or Gaussian with scale $c/ε$). The output is reference-relative and is \emph{not} a formal differential-privacy estimator. We formalize BUA/TDA over hierarchical feature structures and validate the framework on object detectors (YOLO, PPDPTS, DETR) and on the \emph{visual encoders} of VLMs (CLIP, LLaVA, BLIP). BUA and TDA yield comparable deviation trends; EMPA provides a stable alignment signal under the reported setups. We compare with generic discrepancy baselines (Chi-square, K-L, MMD) and with task-relevant baselines (MomentReg, NoiseMLE, Wass-1). Results are reported as mean$\pm$std over multiple seeds with confidence intervals in the supplementary materials. This work contributes a learnable, interpretable modeling perspective for privacy-aligned hierarchical representations rather than a post hoc audit only. Source code: \href{https://github.com/mabo1215/bodhi-vlm.git}{Bodhi-VLM GitHub repository}
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。