Cryptographic Building Block

Cryptographic building blocks are fundamental components enabling secure computation on sensitive data, particularly within machine learning. Current research focuses on improving the efficiency of these blocks, especially within homomorphic encryption schemes and secure multi-party computation, often involving optimizations for specific model architectures like neural networks with reduced ReLU operations or tailored skip connections. This work aims to enhance the practicality of privacy-preserving machine learning by reducing computational overhead and improving the performance of encrypted inference, impacting both data security and the development of trustworthy AI systems.