Bodhi VLM: Privacy-Alignment Modeling for Hierarchical Visual Representations in Vision Backbones and VLM Encoders via Bottom-Up and Top-Down Feature Search
This work addresses privacy modeling for hierarchical visual representations in vision backbones and VLM encoders, offering an interpretable framework, though it appears incremental as it builds on existing clustering and assessment methods.
The authors tackled the problem of modeling how noise perturbations in hierarchical visual representations align with a declared privacy budget, proposing Bodhi VLM, a framework that links sensitive concepts to layer-wise features and uses bottom-up and top-down strategies to locate sensitive regions, validated on object detectors and VLM encoders with comparable deviation trends and stable alignment signals.
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}