HomaVo

HomaVo: Open-Source Homomorphic Encryption Library for Privacy-Preserving Machine Learning

HomaVo is an open-source homomorphic encryption software library focused on enabling privacy-preserving machine learning. It allows computations on encrypted data without access to decryption keys.

What is HomaVo?

HomaVo is an open-source homomorphic encryption library developed by researchers at IBM. It implements various homomorphic encryption schemes that allow computations directly on encrypted data without requiring access to decryption keys.

Homomorphic encryption enables privacy-preserving machine learning by letting machine learning models train on encrypted data. HomaVo makes it easier for developers to build applications using homomorphic encryption for confidential computing use cases like financial analytics, healthcare research, and more.

Key capabilities and features of HomaVo include:

• Implemented homomorphic encryption schemes including BFV, CKKS, and TFHE
• GPU acceleration support for faster homomorphic computations
• Benchmarking tools for evaluation and performance measurements
• Reference implementations of privacy-preserving ML models
• Easy integration into applications via API
• Active development community on GitHub

HomaVo is licensed under the Apache 2.0 open source license. It continues to evolve with state-of-the-art homomorphic encryption research to make confidential computing more practical and widely adopted.

HomaVo Features

Official Links