Google Expands Vertex Managed AI Platform

Google has announced new capabilities for Vertex AI, its managed AI platform designed to help companies accelerate the deployment of AI within their organizations, also known as MLOps.

The various features and improvements to Vertex AI fit into a four-part framework that Google developed in its discussions with customers and partners, says Google in a blog post last week.

Accelerating ML deployment in production

For a start, a partnership with Nvidia made it possible for data scientists to start model development in one click by leveraging Nvidia AI software solutions, simplifying the deployment of Jupyter Notebooks from over 12 complex steps to a single click.

In that vein, the new Vertex AI Training Reduction Server is a training reduction server that makes it possible to the performance of multi-node distributed training on Nvidia GPUs to significantly reduce the training time required for large language workloads.

Google also announced the preview of Vertex AI Tabular Workflows, which includes a glass box and managed AutoML pipeline that lets users observe and interpret each step in the model building and deployment process.

Other notable improvements include:

• Support for new managed algorithms including advanced research models like TabNet (a deep learning model for tabular learning).
• A data partnership with Neo4j for data scientists to explore, analyze, and engineer features from connected data in Neo4j to models with Vertex AI.
• Preview of Serverless Spark for data scientists to launch a serverless spark session within their notebooks.
• Preview of Vertex AI Example-based Explanations to help data scientists identify mislabeled examples in their training data.

Referring to the preview of the Vertex AI Examples-based Explanations, Andrew Moore, a vice president and general manager of Google Cloud AI noted that data scientists shouldn’t need to be infrastructure engineers or operations engineers to keep AI models accurate, explainable, scaled, and disaster-resistant.

With the ability to access example-based explanations to quickly diagnose and treat issues, data scientists can now maintain a high bar on model quality by quickly identifying errors or determining what data to collect to improve model accuracy, Moore noted.

“The new Vertex AI features we’re launching today will continue to accelerate the deployment of machine learning models across organizations and democratize AI so more people can deploy models in production, continuously monitor and drive business impact with AI,” wrote Henry Tappen, a Google Cloud group product manager in an email.

Image credit: iStockphoto/elenabs