Building Your Own AI System

Google is one of the biggest tech companies paving the way for artificial intelligence and machine learning, and a recent announcement from the company stands to bolster that reputation. This week, Google announced the launch of a new service that will enable both businesses and individuals to begin building their own AI systems.

Officially called Google Cloud AutoML, the service comes in the wake of Google's recognition that only a handful of big businesses currently have the budgets necessary to take advantage of AI and machine learning. At the same time, these are often the businesses best positioned to bring on new talent specializing in AI and machine learning engineering. While Google does have pre-trained models, they're typically trained to perform very specific tasks. This programming also limits their recognition to only the objects they've been previously trained to recognize.

In a blog post, Google explained how Cloud AutoML will help businesses design their own customizable AI systems using advanced techniques like learning2learn and transfer learning. The first release of Cloud AutoML will be Cloud AutoML Vision, a service that focuses on making it easier to build machine learning models for image recognition.

We believe Cloud AutoML will make AI experts even more productive, advance new fields in AI and help less-skilled engineers build powerful AI systems they previously only dreamed of. While we’re still at the beginning of our journey to make AI more accessible, we’ve been deeply inspired by what our 10,000+ customers using Cloud AI products have been able to achieve.

AI Expertise

One of the biggest takeaways from the Cloud AutoML announcement is how it can facilitate the expansion of a dwindling talent pool. As said before, big businesses have the best chance of recruiting machine learning experts, but there's a scarce amount of such talent available.

According to MIT Technology Review, chief scientist at Google Cloud Fei-Fei Li acknowledged the gap that exists between developers who have access to AI tools and those who do not.

“We need to scale AI out to more people,” Li said. “But there are an estimated 21 million developers worldwide today. We want to reach out to them all, and make AI accessible to these developers.”

There's also the not-so-small matter of the computing power needed to run an AI system. Recognizing that smaller businesses and individuals may not be capable of supporting such systems, Google aims to shoulder the burden through the use of Cloud AutoML.

With Cloud AutoML Vision specifically, developers will be able to utilize their systems directly on Google Cloud. In the event the images being used to train a system don't have labels of their own, developers can rely on Google's team of in-house labelers to "review your custom instructions and classify your images accordingly." With Google taking care of the computational demand, problems could still crop up if a large number of users decide they want to push the capabilities of the service.

Cloud AutoML Vision is only the first of several Cloud AutoML projects Google has in development, as the company has plans to provide AI developers with even more ways to engage in machine learning development. Perhaps, with more people understanding both technologies, we can find a better way to incorporate both into our lives without allowing machines to completely take over.

Share This Article