Gretel and Google Cloud partner on synthetic data

Gretel and Google Cloud harness the power of synthetic data to accelerate safer adoption of generative AI in the enterprise.
Gretel partners with Google Cloud to harness the power of synthetic data and accelerate adoption of safer generative AI in the enterprise.

We’re excited to announce the availability of Gretel’s synthetic data platform for Vertex AI and BigQuery developers, as well as Gretel’s launch in the Google Cloud Marketplace. 

With Vertex AI, MLOps teams have all the functions they need for rapid and scalable AI development at their fingertips. What's often missing, though, is a steady supply of quality data to train and deploy resilient ML models that perform well on downstream tasks. Improving model accuracy requires robust training datasets that cover edge cases and that account for changing real-world dynamics that otherwise lead to problems like data and concept drift. However, the only way to get this improved accuracy is to leverage large volumes of sensitive data, which presents security and privacy risks. With Gretel’s models, which are available in multiple modalities like image, tabular, text, time-series, and relational, developers can create secure versions of sensitive data from existing Google Cloud storage buckets and deliver them directly to Vertex. This approach addresses data compliance concerns and any data supply constraints, as you can access and create whatever data you need, on-demand.

Google Cloud’s partnership with Gretel makes it simple to generate anonymized and safe synthetic data for enterprises blocked by data sharing limitations or a lack of data. Synthetic data also enables developers to experiment with, fine-tune, and operationalize foundation models without giving these models access to their raw and sensitive data. This unlocks “the last mile” in generative AI, where the data that’s most valuable to train a model isn’t available in the public domain or simply can’t be shared. Developers that use Vertex AI for MLOps will now be able to use synthetic data to meet privacy and data augmentation requirements, resulting in accelerated ML R&D and operations.

The entire collection of Gretel’s models is available for Google Cloud and Vertex customers to incorporate into their services, including APIs to:

  • Generate synthetic data to use in place of production to enable safe data sharing.
  • Generate synthetic data to augment training data when existing training data isn't sufficient.
  • Generate synthetic data when there's no existing data to work with (the “cold start” problem) using Gretel’s generative machine learning models. 

Research has shown that synthetic data can be as good or even better than real-world data for data analysis and training AI models; and that it can be engineered to reduce biases in datasets and protect the privacy of any personal data that it’s trained on. With the right tools, synthetic data is also easy to generate, so it is considered a fast, cost-effective data augmentation technique.

“Gretel’s platform provides important tooling for developers utilizing data and building generative AI models and applications,” said Manvinder Singh, Managing Director, Partnerships at Google Cloud. “By integrating with Vertex AI and bringing its platform to Google Cloud Marketplace, Gretel provides more choice and capability for customers to both deploy Gretel on Google Cloud and to utilize its capabilities as they innovate with generative AI on Google Cloud.”

If you’re reading this, you know that the promise of generative AI is unquestionable, and that enterprise leaders everywhere are imploring their teams to incorporate ML and AI into their R&D and products. The biggest hurdle to jumping in, in addition to having limited data to start with, is ensuring the privacy of that data and the safety of the models being used. Safe incorporation of generative AI means ensuring that what a business gets out of its generative model is exactly what it expects from its data. By using Gretel’s synthetic data, with its pre- and post-validation ensuring the privacy of the model and the utility of the data, enterprise leaders seeking to win in a world being revolutionized by generative AI can do so safely.

If you are a Vertex user, give Gretel a spin for free and start building safer models with synthetic data. Going forward, Gretel will also work closely with Google Cloud to accelerate deployment and availability within the GCP Marketplace, including adding integration with BigQuery for fully automated workflows. Want to use Gretel and Vertex AI within your Google Cloud deployment? Sign up here for early access.