This article was first published on our sister Site, Whats New On The Net.
With Vertex AI, Google wants to push artificial intelligence (AI) & the machine learning (ML) it relies on into the mainstream. Google maintains that its new AI solution needs 80% less code input than their competitor’s to obtain results. It’s a managed platform, which Google hopes will narrow the gap between experimentation & production environments that in the past has cost dabblers high expenditure with inconsistent outcomes.
The Cloud service offered by Google for building ML is collected under one unified UI & API, giving users the opportunity to deploy AI solutions at scale in a single, unified environment, which means that data science & ML engineering teams can:
- Access Google’s internal AI toolkit – computer vision, language, conversation & structured data, which is continuously updated by Google’s latest research
- Build useful Al models much faster – deploy Machine Learning Operations (MLOPs) features such as Vertex Vizier (increases the rate of experimentation), the Vertex Feature Store to serve, share, & reuse ML features, Vertex Experiments to accelerate model selection & Vertex ML Edge Manager to monitor models via their API
- Effectively manage models – Vertex Model Monitoring, Vertex ML Metadata & Vertex Pipelines assist in managing complexity & repetition
An example of Vertex in action is L’Oréal’s ModiFace, which allows consumers to try on beauty products virtually. L’Oréal is using Vertex to train its ML models using thousands of images & other vital information. It’s a production scale investment by L’Oréal that is paying dividends & leads the way for other like minded companies to follow.
Generally, once developed ML models are static, but with Vertex AI, developers & data analysts can keep models current by regularly updating them to meet fast changing needs.
“Vertex AI is a single platform with every tool you need, allowing you to manage your data, prototype, experiment, deploy models, interpret models, & monitor them in production without requiring formal ML training.”
This also means that highly trained developers will be less essential & that ordinary applications can be developed to take advantage of ML for most every business.
Google has a best-practices guide on its Cloud Architecture Center to help onboard newbies & familiarize them with the suite of tools Google suggests they employ to get the most out of Vertex AI – such as BigQuery, which assists in managing huge sets of data quickly & efficiently.
Google has offered Interested parties an opportunity to sign up to attend the “Applied ML Summit” for data scientists & ML engineers, which will showcase Vertex AI & will be held on June 10 this year. They also suggest investigating the resources offered by Accenture & Deloitte who have created design workshops, proof of value projects, & operational pilots to get developers & companies started on groundbreaking ML technology.