With around 20 million developers worldwide and only 2 million data scientists, Google Cloud is aiming to reduce the burden and make AI faster, simpler and more useful.
Google Cloud has unveiled two tools to make it easier for businesses to use machine learning in their processes and overcome the “daunting” prospect of fully embracing AI.
Kubeflow Pipelines provides a “workbench” to compose machine learning (ML) workflows, and packages ML code to make it reusable to other users across an organisation.
The tool, open now on GitHub, allows multiple members of a team to collaborate on an ML application by chipping in with data or an API.
Data scientists can also test several ML techniques using the tool, to see which one works best for them and their company’s application.
Kubeflow Pipelines is a new component of Kubeflow, Kubernetes’ open source project.
AI Hub, meanwhile, is a one-stop destination for professionals to upload and share machine learning resources for sharing within their own organisation.
This makes it easy for businesses to reuse pipelines and deploy them to production in Google Cloud Platform or on hybrid infrastructure using the Kubeflow Pipeline system.
It compiles Google’s machine learning tools, including Google Cloud AI and Google Research, in the one place. Once Google Cloud AI Hub reaches beta, it will expand to include partner solutions.
Google Cloud said that organisations including Nvidia, Cisco, and Intel are among the companies working closely to adopt Kubeflow Pipelines.
Google Cloud AI Hub: One-Stop Destination for Plug-and-Play ML Content
Kaustubh Das, VP for data centre product management at Cisco, said: “Cisco’s significant contributions to Kubeflow aims to simplify hybrid/multi cloud AI/ML lifecycle management”, adding that Pipeline “promises a radical simplification of ML workflows which are critical for mainstream adoption”.
Nvidia is currently integrating RAPIDS, its new suite of open source data libraries, into Kubeflow.
Google Cloud said it hopes to address a lack of machine learning knowledge in the workforce that makes it challenging to build AI resources, part of its AI principles.
“Our goal is to put AI in reach of all businesses,” the company said in a blog post. “But doing that means lowering the barriers to entry.
“That’s why we build all our AI offerings with three ideas in mind: Make them simple, so more enterprises can adopt them; make them useful to the widest range of oragnisations; and make them fast, so businesses can iterate and succeed more quickly.”
Along these lines, earlier this year Google Cloud announced AutoML Vision, allowing businesses and developers to create custom machine learning models for image recognition.
It also has Advanced Solutions Labs in five physical locations – Sunnyvale, New York, Dublin, Singapore, and Tokyo – for on-site collaboration with Google’s machine learning engineers.