Privacy-preserving datasets by capturing feature distributions with Conditional VAEs

TL;DR

This paper presents a novel privacy-preserving data generation method using Conditional Variational Autoencoders trained on features from foundation models, enhancing data diversity and privacy in sensitive domains like medicine.

Contribution

Introduces a CVAE-based approach trained on foundation model features to generate diverse, privacy-preserving synthetic data, outperforming traditional anonymization methods.

Findings

Outperforms traditional anonymization in data diversity

Provides higher robustness against perturbations

Generates unbounded, privacy-respecting synthetic datasets

Abstract

Large and well-annotated datasets are essential for advancing deep learning applications, however often costly or impossible to obtain by a single entity. In many areas, including the medical domain, approaches relying on data sharing have become critical to address those challenges. While effective in increasing dataset size and diversity, data sharing raises significant privacy concerns. Commonly employed anonymization methods based on the k-anonymity paradigm often fail to preserve data diversity, affecting model robustness. This work introduces a novel approach using Conditional Variational Autoencoders (CVAEs) trained on feature vectors extracted from large pre-trained vision foundation models. Foundation models effectively detect and represent complex patterns across diverse domains, allowing the CVAE to faithfully capture the embedding space of a given data distribution to generate (sample) a diverse, privacy-respecting, and potentially unbounded set of synthetic feature vectors. Our method notably outperforms traditional approaches in both medical and natural image domains, exhibiting greater dataset diversity and higher robustness against perturbations while preserving sample privacy. These results underscore the potential of generative models to significantly impact deep learning applications in data-scarce and privacy-sensitive environments. The source code is available at https://github.com/francescodisalvo05/cvae-anonymization .