Elevate your enterprise data technology and strategy in Transform 2021.
During a virtual keynote at Google I/O 2021, Google’s developer conference, Google announced the launch generally availability of Vertex AI, a controlled AI platform. It is designed to help businesses to accelerate the deployment and maintenance of AI versions, Google says, by requiring nearly 80percent fewer lines of code to train a version versus aggressive platforms.
Data scientists often grapple with the challenge of piecing together AI solutions, developing a lag time in model experimentation and development. In a recent Alation report, the majority of respondents (87%) pegged data quality problems as the reason their associations failed to execute AI. That’s why companies enjoy Markets and Markets anticipate that the data recovery industry, including companies that offer data cataloging and curation tools, will probably be worth upwards of $3.9 billion from the end of 2021.
To tackle the challenges, Vertex brings together Google Cloud solutions for AI beneath a unified UI and API. Vertex lets customers build, train, and deploy system learning models in a single environment, moving models from experimentation to manufacturing while discovering patterns and anomalies and making forecasts.
“Vertex was designed to help customers with four things,” Google Cloud AI product management director Craig Wiley told VentureBeat in an interview. “The first is, we want to help them increase the velocity of the machine learning models that they’re building and deploying. Number two is, we want to make sure that they have Google’s best-in-class capabilities available to them. Number three is, we want these workflows to be highly scalable. … And then number four is, we want to make sure they have everything they need for appropriate model management and governance.
“Ultimately, the aim here is to figure out how we could accelerate companies discovering ROI with their machine learning.”
Fully managed AI
Vertex offers access to the MLOps toolkit used internally at Google for computer vision, language, conversation, and structured data workloads. MLOps, a compound of “machine learning” and “information technology operations,” is a newer discipline involving collaboration between data scientists and IT professionals with the aim of productizing machine learning algorithms.
Vertex’s other headlining features include Vertex Vizier, which aims to increase the rate of experimentation; Vertex Feature Store, which lets practitioners serve, share, and reuse machine learning features; and Vertex Experiments, which helps with model selection. There’s also Vertex Continuous Monitoring and Vertex Pipelines, which support self-service model maintenance and repeatability.
Customers including L’Oréal-owned ModiFace and Essence are using Vertex for production models, Google says. According to Jeff Houghton, ModiFace’s COO, Vertex allowed the company to create augmented reality technology”incredibly close to actually trying the product in actual life.” As for Essence, SVP Mark Bulling says that Vertex is enabling its data scientists to quickly create new models based on changes in environments while also maintaining existing models.
“Once your model’s in production, the world is constantly changing, and so the accuracy of these models is continuously degrading over time. You have to keep tabs on your model and know how it’s performing, and be ready to respond if it