Criptografia homomórfica no aprendizado de máquina.

Abstract

Homomorphic encryption represents a paradigm shift in the realm of secure data processing, allowing computations on encrypted data without the need for decryption. This capability promises significant advancements in enhancing privacy and security across various domains, including cloud computing, healthcare, finance, and also in machine learning. This Final paper delves into the fundamentals of homomorphic encryption in machine learning, elucidating its mathematical underpinnings and exploring its practical applications. Through a comprehensive review of existing literature and methodologies, this research evaluates the strengths, weaknesses, and potential challenges associated with it. Additionally, it investigates the performance implications and computational overhead incurred by different homomorphic encryption schemes. The study also examines real-world use cases and implementation scenarios to assess the viability and effectiveness of homomorphic encryption for secure data processing and privacy-preserving technologies.