Google Launches BigQuery ML To Bring Machine Learning to More Businesses
- By Richard Seeley
Business adoption of artificial intelligence (AI) is happening at 60 percent of companies, according to Google, but now the company is seeking to reach the other 40 percent with the launch of BigQuery ML and updates of other data analytics tools.
Lack of expertise, especially that rare breed known as data scientists, is a roadblock for many businesses looking at adopting AI and machine learning (ML), according to Rajen Sheth, senior director of product management, Android and Chrome for Business and Education at Google.
"For many businesses, there are significant hurdles to building the analytics pipeline necessary for AI," he wrote in a recent announcement of Google's data analytics initiative. "Creating a team of in-house data scientists is impractical for many. Data analysts, typically trained in SQL, aren't always familiar with the processes and programming languages used for machine learning. And workstreams that involve moving data out of an enterprise data warehouse can be time-consuming and costly."
BigQuery ML, now in beta, is designed to break down the barriers by providing a tool that does not require a data scientist to operate it, Sheth said. Google is also updating other products in its data analytics and Cloud AI services offering.
"BigQuery ML puts the power of predictive analytics in reach of millions of users -- even those without a data science background," he wrote. "By bringing ML to where customers already store their data, BigQuery ML helps customers quickly create and deploy models, accelerating speed to market. They can run models at scale on large datasets. And they can do it all using simple SQL commands."
Google provides an in-depth look at BigQuery ML on its data analytics blog.
An updated version 0.2 of Google's Kubeflow, originally announced in 2017, is also included in the company's initiative to simplify AI and ML for late adopters in the business community.
Continuing Google's theme of simplifying AI, it said Kubeflow is designed to make it easier to work with machine learning Kubernetes software stacks, including TensorFlow and Scikit-Learn. Kubeflow v0.2 offers an upgraded user interface for navigating among components, as well as improvements for monitoring and reporting, the company said. Details on the updated tool are available on Getting Started with Kubeflow.
Google also announced alpha versions of Cloud TPU v3 and Cloud TPU Pod, part of its AI infrastructure providing custom ASICs designed to power machine learning workloads "from the cloud to the edge."
The company says each TPU delivers as much as 180 teraflops of floating-point performance. TPU pods allow TPUs to operate in groups for more compute power. The alpha version of the Cloud TPU Pod provides up to 11.5 petaflops, Google said.
The company's announcement included a testimonial for Cloud TPU Pods from eBay: "Cloud TPU Pods have transformed our approach to visual shopping by delivering a 10X speedup over our previous infrastructure," says Larry Colagiovanni, vice president of new product development at eBay. "We used to spend months training a single image recognition model, whereas now we can train much more accurate models in a few days on Cloud TPU Pods. We've also been able to take advantage of the additional memory the TPU Pods have, allowing us to process many more images at a time. This rapid turnaround time enables us to iterate faster and deliver improved experiences for both eBay customers and sellers."