Federated learning enables private tabular data to image encoding
Fed-IGTD: A federated learning-based framework for tabular-to-image encoding
Machine LearningComputer Vision and Pattern Recognition
Summary
Tabular data, like spreadsheets, can be turned into picture-like forms for making predictions, but doing this usually means gathering all data in one place, which risks privacy. The authors designed Fed-IGTD, a way for many users to jointly train models without sharing raw data, by only sending some statistical summaries instead. They tested privacy protections to see how they affect prediction quality, finding that hiding the mapping step keeps accuracy high but protecting the training with noise hurts results when users are few. The research helps balance privacy and usefulness in data that is spread across users.
What this means in practice
- •For healthcare data teams: Enable joint model training on private patient data stored in separate hospitals while keeping individual records confidential using Fed-IGTD.
- •For financial fraud analysts: Train fraud detection models across multiple banks’ customer data without sharing sensitive financial records by aggregating statistical features as Fed-IGTD does.
Authors
David Chushig-Muzo, Eva Milara, Francisco J. Lara Abelenda, Luis Carlos Estrada-Petrocelli, Diego H. Peluffo-Ordóñez
Abstract
Tabular-to-image encoding methods, which transform tabular data into image-like representations, have shown promising results in predictive tasks. Despite their benefits, these methods follow a centralized paradigm, in which the data are collected and models are trained on a central device. This raises privacy and security concerns, particularly the risk of data leakage, unauthorized data collection and access to users' private information. Federated Learning (FL) provides a decentralized framework in which multiple clients jointly optimize a shared model without directly sharing training data. We introduce Fed-IGTD, a horizontal FL-based approach in which clients share a common feature space and upload only local statistics, from which the server reconstructs the global correlation matrix and obtains a feature-to-pixel mapping to train a convolutional neural network via Federated Averaging (FedAvg). We studied two differential privacy (DP) mechanisms: (i) a one-shot Gaussian mechanism protecting the correlation-aggregation step; and (ii) a client-level DP-FedAvg variant protecting the training with privacy loss composed via a Renyi-DP accountant. We evaluated Fed-IGTD on five datasets under a systematic sweep of non-IID severity and privacy budgets. Privacy protection applied to the mapping-construction step has little impact on predictive performance. By contrast, DP-FedAvg substantially degrades utility at moderate privacy budgets with a small client cohort (K=5). Ablation study indicates that this degradation is driven mainly by the injected noise rather than by uniform averaging or update clipping. Macro-F1 exhibits sensitivity to performance degradation than accuracy under non-IID heterogeneity and class imbalance.