At Google I/O, we announced the full launch of Vertex AI, a manageable machine learning (ML) platform that allows enterprises to accelerate deployment and maintenance of artificial intelligence (AI) models. Vertex AI reduces the number of lines of code required to train a model by nearly 80% compared to competing platforms, 1 enabling data scientists and ML engineers of all professional levels to implement machine learning operations (MLOps) to effectively set up and manage ML projects throughout the development cycle.
Today, data scientists struggle with the challenge of manually piecing together ML point solutions, creating lag times in model development and experimentation, resulting in few models ever making it into production. To address these challenges, Vertex AI brought together Google Cloud services for building ML under a unified user interface and API to simplify the process of building, training, and deploying machine learning models on a large scale. In this single environment, customers can move models from experimentation to production more quickly, detect patterns and anomalies more effectively, make better predictions and decisions, and are generally more flexible in the face of changing market dynamics.
Through Google’s decades of innovation and strategic investments in AI, the company has learned important lessons about how to build, deploy, and maintain ML models in production. These insights and engineering have been built into the foundation and design of Vertex AI, and will continue to be fleshed out by new innovations at Google Research. Now, through Vertex AI, data science and ML engineering teams can for the first time.
Get access to the AI toolkit for internal use at Google, including computer vision, language, conversation and structured data, and continuously enhanced by Google Research.
Leverage new MLOps features such as Vertex Vizier to increase the speed of Experiments, fully managed Vertex Feature Store to help practitioners service, share and reuse ML features, and Vertex Experiments with faster model selection, Accelerate model deployment into production. If your data needs to stay on the device or in the field, Vertex ML Edge Manager can deploy and monitor models on the Edge with automated processes and flexible apis.
Confidently manage models to simplify end-to-end ML workflows by eliminating the complexity of self-service Model maintenance and repeatability through MLOps tools such as Vertex Model Monitoring, Vertex ML Metadata, and Vertex Pipelines.
“In establishing vertex of artificial intelligence, we have two guidelines: let the data out of the coordination of weeds, scientists and engineers to create a whole industry shift, make everyone is serious about artificial intelligence from pilot transferred to full production in purgatory,” Google vice President and cloud cloud, general manager of artificial intelligence and industry solution Andrew Moore said.” “We are very proud of our results on this platform as it enables serious deployment of a new generation of AI that will enable data scientists and engineers to do meaningful and creative work.”
“Enterprise data science practitioners looking to apply AI to the entire enterprise are not looking for a tool arrangement. Instead, they wanted tools that tamed the ML lifecycle. “Unfortunately, this is not a minor issue,” says Bradley Shimmin, principal analyst for AI platforms, analytics and data management at Omdia. “This requires a supportive infrastructure that unifies the user experience, puts AI itself as a supportive guide, and puts data at the heart of the process — while encouraging flexible adoption of technologies.”
ModiFace uses Vertex AI to revolutionize the beauty industry
ModiFace, part of L ‘Oreal, is the global market leader in augmented reality and ARTIFICIAL intelligence for the beauty industry. ModiFace has created new services for consumers that enable them to try out beauty products such as hair color, makeup and nail color in real time. ModiFace is using vertex AI platform to train AI models for all its new services. For example, ModiFace’s skin diagnostics are trained on thousands of images from L ‘Oreal Research & Innovation, the company’s dedicated research department. The service combines L ‘Oreal’s scientific research with ModiFace’s ARTIFICIAL intelligence algorithm to enable people to obtain highly accurate customized skin care routines.
“We provide people with an immersive, personalized experience that lets them buy in confidence, whether it’s a virtual try on at online checkout or helping to understand what brands are right for everyone,” said Jeff Houghton, chief operating officer of ModiFace, which is owned by L ‘Oreal. “As more and more of our users look for information at home, on their phones, or at any other point of contact, Vertex AI allows us to create technology that comes incredibly close to actually trying out products in real life.”
Essence was born for the age of algorithms with the help of Vertex AI
Essence, a global data and measure-driven media organization affiliated with WPP, is expanding the value of AI models made by its data scientists by integrating their data scientist workflows with developers using Vertex AI. Historically, the AI models created by data scientists didn’t change once they were created, but as human behavior and channel content continues to change, so does the way they operate as the digital world evolves. With Vertex AI, developers and data analysts can regularly update models to meet these rapidly changing business needs.
“At Bejet, we are measured by our ability to keep up with the rapidly evolving needs of our customers,” said Mark Bulling, Senior vice president of product innovation at Bejet. Vertex AI gives our data scientists the ability to quickly create new models in response to changes in the environment, while also allowing our developers and data analysts to maintain models for easy expansion and innovation. “The MLOps capabilities in Vertex AI mean we can stay ahead of our customers’ expectations.”
A unified data science and ML platform for all skill levels
MLOps life cycle
One of the biggest challenges we hear from clients is finding talent to work on machine learning projects. Nearly two-fifths of companies cite a lack of technical expertise as a major barrier to using AI technology.
Vertex AI is a single platform with every tool you need, allowing you to manage your data, prototypes, experiments, deploy models, interpret models, and monitor them in production without formal ML training. This means your data scientist doesn’t need to be an ML engineer. With Apex AI, they have the ability to move quickly, but with a safety net, their work can always be jump-started. The platform facilitates responsible deployment and ensures that you move more quickly from test and model management to production and ultimately drive business results.
“Among Sabre’s travel AI technologies, Google’s Vertex AI gives our technologists the tools they need to quickly experiment and deploy smart products across the travel ecosystem. “This progress demonstrates how the power of partnerships between our teams is helping to accelerate Sabre’s vision for the future of personalized travel,” said Sundar Narasimhan, Sabre’s senior vice president and president of Laboratory and Product strategy.
“As Iron Mountain offers our customers more sophisticated technology and digital transformation services, having a comprehensive platform like Vertex AI will allow us to simplify building and running ML pipelines and simplify MLOps for our AI/ML teams,” said Narasimha Goli, Vice President of global Digital Solutions innovation.
Start using vertex ARTIFICIAL intelligence
For more information on how to get started with the platform, check out our ML Best Practices on GCP, the Practitioner’s guide to MLOps white paper, and sign up for our Application ML Summit for data scientists and ML engineers on June 10. We can’t wait to work with you to apply ground-breaking machine learning techniques to enhance your skills, career and business.
To further support Vertex AI’s launch, Accenture and Deloitte have created design workshops, proof-of-value programs, and operational pilots to help you get up and running on the platform.
The original link: cloud.google.com/blog/produc…