Secure and Compliant Cloud-based Healthcare Analytics using Federated Learning and Differential Privacy
Bibliographic record
Abstract
In the era of data-driven healthcare, ensuring the security, privacy, and regulatory compliance of cloud-based artificial intelligence (AI) systems has become critically significant. With sensitive patient data increasingly stored and processed in distributed cloud environments, safeguarding information while maintaining model performance is an initial requirement for ethical and legal AI deployment. Traditional machine learning (ML) models, though effective in terms of accuracy, require raw data transmission and centralized storage, exposing systems to increased risks of data breaches, unauthorized access, and privacy violations. Moreover, these existing systems often lack mechanisms for auditable compliance, adaptive privacy control, and scalable collaboration across multiple institutions. To address these limitations, this study presents a two-stage privacy-preserving AI framework for cloud and edge computing that integrates EfficientNetB0-based feature extraction, compressed variational autoencoders (CVAE) for anomaly detection, differential privacy (DP) transformation, federated learning (FL), and reinforcement learning (RL)-based adaptive privacy enforcement, enabled by blockchain-based compliance logging. Experimental evaluation on the Medical Information Mart for Intensive Care III (MIMIC-III) dataset demonstrates a high prediction accuracy of 97.3%, an AUROC of 0.981, and robust performance across varying privacy budgets (∈) . The model also shows steady convergence across federated training rounds without sacrificing privacy or accuracy. This study provides valuable insights into designing secure, privacy-aware, and compliant AI architectures for healthcare, offering a practical foundation for future systems deployed in sensitive, multi-institutional environments.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.007 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.003 |
| Open science | 0.002 | 0.003 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".