KIPPS: Knowledge infusion in Privacy Preserving Synthetic Data Generation
Anantaa Kotal, Anupam Joshi
TL;DR
KIPPS tackles the challenge of privacy-preserving synthetic data in domains with discrete, non-Gaussian features and limited data by infusing domain knowledge from Knowledge Graphs into GAN-based training. The approach replaces discrete values with domain properties, groups attributes by properties, and enforces domain rules through a conditional training regime and a domain-rule loss, complemented by a differentially private discriminator via DP-SGD. Evaluations across cybersecurity and healthcare datasets demonstrate improved fidelity, utility, and privacy resilience compared to state-of-the-art tabular data generators, indicating a practical balance for data sharing while maintaining strong privacy guarantees. This framework enables realistic, domain-compliant synthetic data with provable privacy, facilitating safer data sharing and collaboration in sensitive domains.
Abstract
The integration of privacy measures, including differential privacy techniques, ensures a provable privacy guarantee for the synthetic data. However, challenges arise for Generative Deep Learning models when tasked with generating realistic data, especially in critical domains such as Cybersecurity and Healthcare. Generative Models optimized for continuous data struggle to model discrete and non-Gaussian features that have domain constraints. Challenges increase when the training datasets are limited and not diverse. In such cases, generative models create synthetic data that repeats sensitive features, which is a privacy risk. Moreover, generative models face difficulties comprehending attribute constraints in specialized domains. This leads to the generation of unrealistic data that impacts downstream accuracy. To address these issues, this paper proposes a novel model, KIPPS, that infuses Domain and Regulatory Knowledge from Knowledge Graphs into Generative Deep Learning models for enhanced Privacy Preserving Synthetic data generation. The novel framework augments the training of generative models with supplementary context about attribute values and enforces domain constraints during training. This added guidance enhances the model's capacity to generate realistic and domain-compliant synthetic data. The proposed model is evaluated on real-world datasets, specifically in the domains of Cybersecurity and Healthcare, where domain constraints and rules add to the complexity of the data. Our experiments evaluate the privacy resilience and downstream accuracy of the model against benchmark methods, demonstrating its effectiveness in addressing the balance between privacy preservation and data accuracy in complex domains.