Google has made its Cloud AutoML service available on Kaggle, the Google-owned online community currently home to 3.5 million data scientists and machine learners.
ML and AI for developers
The platform is used to learn and apply machine learning for businesses and developers looking to advance artificial intelligence. Arguably the community is a step towards closing the skills gap in advanced technologies.
AutoML is a suite of products to help users build custom machine learning models for various tasks such as vision, language or structured data.
Essentially, the tool will ingest data from the user’s own software development kit (SDK) or web user interface (UI) and, with minimal input from the user, will output a trained model that can be deployed to Google’s cloud infrastructure.
On the integration, Devvret Rishi, product manager of Kaggle at Google, says the time efficiency of AutoML is a key factor for what makes the tool useful.
Discussing the Kaggle Days at Cloud Next ‘19, Rishi says, “What especially drew our interest was that the team using AutoML was able to get strong performing results quickly, with low effort and no domain expertise or supervision.
“What especially drew our interest was that the team using AutoML was able to get strong performing results quickly, with low effort and no domain expertise or supervision.”
The new integration enables Kaggle users to use AutoML SDK within Jupyter Notebooks, the open-source web applications used by data scientists to create and share documents with live code, equations, visualisations and narrative text.
Rishi says Google plans to make AutoML more accessible to users of Kaggle. It will do this by offering GCP credits throughout the year to subsidise the costs of using the service, and give new Google accounts that sign up for GCP $300.