Google has announced its new Cloud AutoML in its alpha phase. That means that the service is still being tested internally and won’t be for release just yet. However, for those who want to build their own custom models from it, that is already good news as it means that the service is on its way to being released in the future.
For people who want to use it for their own programs, Cloud AutoML can have advanced techniques to machine learning, including learning2learn and transfer learning. Learning2learn is a technique used for automated machine learning. Meanwhile, transfer learning is when a fully trained model is used to set out for new categories to learn.
Google Bringing Machine Learning to The Masses
Cloud AutoML is based on image recognition processes. This means that more accurate models could be made through it. Machine learning is just one of the many models for artificial intelligence, or AI. AI is seen to be the next technological advancement which would take computers, robots and others to be independent and learn on their own.
With Cloud AutoML, even those with limited knowledge in AI can build an application or model in just minutes. This is in part because it has a simple graphical user interface (GUI) which makes it easy for users to navigate the environment. Cloud AutoML has been the result of a collaboration between Google Brain and other Google AI teams.
Ultimately the goal is to have Cloud AutoML be available to many people. The first step is to of course test it, and with that it is entering its alpha testing stage before opening it up for more users to test it.
“This is the first step towards the non-technical end user,” Holger Mueller, Principal Analyst at Constellation Research
With AI moving towards a more mass market approach, would it really become the next big advancement in technology? Comment on the section below.