Privacy-Preserving Data Leakage Detection Using Homomorphic Encryption and Deep Learning
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Data leakage poses significant threats to organizational security, with traditional detection methods requiringaccess to sensitive information in plaintext. This paper presents a novel framework for privacy-preserving data leakagedetection that combines fully homomorphic encryption (FHE) with deep learning architectures. Our approach enablesorganizations to detect sensitive data exposure without decrypting the monitored content, addressing critical privacyconcerns. We introduce HE-DLDNet, a modified transformer-based neural network that operates directly on encrypted datastreams, achieving 94.3% detection accuracy while maintaining computational efficiency. Experimental results onbenchmark datasets demonstrate that our method reduces computational overhead by 73% compared to existinghomomorphic approaches while preserving privacy guarantees. The framework supports real-time detection with latencyunder 250ms, making it suitable for production deployment. This work bridges the gap between strong privacy preservationand practical data leakage detection systems.
Fields of Study & Subjects
Privacy-Preserving Technologies in Data
Cryptography and Data Security
Internet Traffic Analysis and Secure E-voting
Publication details
- Journal / source
- Journal of Science Engineering Technology and Management Sciences
- Publication date
- 23 March 2026
- Language
- EN