August 2022 Summaries

Gretel is changing its license for some open-source components to the Gretel Source Available License (SAL). This new license allows users to freely install and incorporate Gretel software packages into their own solutions, but they cannot redistribute them or build a SaaS offering with the packages. The company will also add a Contributor License Agreement (CLA) for contributions made to affected software packages. These changes are aimed at ensuring long-term growth and innovation in the synthetic data community while maintaining a sustainable business model.

The gretel-trainer is an interface designed to be the simplest way to generate synthetic data, offering a preview of new features coming to Gretel's production SDK & API. It enables users to create synthetic data for datasets with millions of rows and thousands of columns without needing to set up or manage infrastructure and GPUs. The gretel-trainer uses Gretel's fully managed cloud service for model training and generation, making it easy to generate state-of-the-art synthetic data. Users can customize their synthetic models by selecting an underlying model and tuning model parameters.

Data sharing is crucial in modern businesses, but it comes with privacy risks. Synthetic data can enable secure data sharing by reducing the risk of linkage attacks that compromise sensitive information. In this blog post, a hypothetical scenario demonstrates how synthetic data can be used to build predictive machine learning models while protecting customer privacy. The use of synthetic data generated through models like CTGAN significantly reduces the likelihood of linkage attacks and provides high-quality statistical patterns for downstream machine learning tasks.

Gretel's Applied Science team, consisting of senior applied scientist Lipika Ramaswamy and principal machine learning scientist Amy Steier, has been nominated for VentureBeat’s Women in AI Awards. Both scientists shared their experiences working on state-of-the-art generative AI models and offered advice to aspiring data scientists. They emphasized the importance of continuous learning, passion for work, and awareness of the environment they are building models in or serving with their technology. The team is excited about the future of responsible AI and its potential impact across various industries.