August 2022 Summaries

The blog post explores a Privacy-Preserving Machine Learning (PPML) solution to the Titanic challenge on Kaggle using the Concrete-ML toolkit, demonstrating the application of Fully Homomorphic Encryption (FHE) to protect data during machine learning predictions without compromising performance. It employs an XGBoost classifier model, which is trained using scikit-learn's GridSearchCV for parameter optimization, and then compares it to a Concrete-ML version that uses FHE. The preparation involves data cleaning, feature engineering, and dummification of categorical variables to transform them into numerical data. Both models perform similarly, with the FHE model achieving a slightly higher accuracy of 78% compared to the XGBoost model's 77%, showcasing the effectiveness of FHE in secure data predictions. The blog emphasizes the ease of integrating Concrete-ML into existing data science workflows without requiring expertise in cryptography, highlighting the potential of FHE for secure machine learning applications.

In July 2022, Zama introduced a new version of the Concrete Framework designed to simplify the application of Fully Homomorphic Encryption (FHE), enabling computations over encrypted data. The blog post illustrates the framework's capabilities by upgrading the implementation of Conway's Game of Life using concrete-boolean to leverage Concrete v0.2's new features. The Concrete Framework, composed of various Rust crates, allows for homomorphic computations using Boolean and integer types, offering an interface similar to Rust for ease of use. The migration from concrete-boolean to Concrete involves straightforward modifications, making code more readable and efficient. The post also details transitioning from homomorphic Boolean to integer types for more intuitive computations, enhancing performance by a factor of 3.3. The performance comparison highlights that while using FheUint2 improves speed, it results in a larger key size. The blog post concludes with plans for future parallelization enhancements and encourages readers to engage with Zama's community through various channels.

Zama, a startup focused on cryptographic solutions, advocates for the principle of transparency in security through open-source practices, opposing security by obscurity. The company is dedicated to making the internet safer by encrypting all data end-to-end and believes that open-source approaches can enhance software security, foster community-driven innovation, and build customer trust. Zama argues that open-source and monetization are not mutually exclusive, citing successful models from industry giants like MongoDB and Elastic Search that combine revenue streams such as support, hosting, proprietary features, and dual licensing. Zama has chosen the BSD3-clear dual licensing model to allow free non-commercial use of its software while monetizing through commercial licenses, enabling the company to remain independent and invest in privacy-preserving technologies. The company encourages collaboration and aims to influence a shift towards more open-source adoption in critical infrastructures to reduce risks associated with unaudited, opaque code.