Google’s AutoML Creates Machine Learning Models Without Programming Experience
Interest in machine learning has exploded in recent years as companies realize it has applications in photography, self-driving cars, games, and more. We’ve reached a point that there aren’t enough experienced programmers and data scientists with the necessary expertise to build these systems. Google’s solution is Cloud AutoML, a point-and-click system for building machine learning models without any coding experience.
Google has long offered pre-trained neural networks accessible via APIs that can perform certain tasks, but that’s only useful if you need exactly what that model does. The gist of Cloud AutoML is that almost anyone can bring a catalog of images, import tags for the images, and create a functional machine learning model based on that. Google does all the heavy lifting behind the scenes, so the customer doesn’t need to know anything about the intricacies of neural network design.
AutoML won’t compete with the cutting edge, highly tuned AI systems an experienced engineer could build, but few businesses have the money or resources to support the development of completely custom machine learning models. AutoML uses a simple graphical interface, allowing the user to drag in a set of images. Then, the platform needs to know how to describe those images. Google does its magic, and you end up with a model running in the cloud that can identify the specified terms in photos. AutoML provides stats on the strength of the model, so you can train it with more data or test with new images.
The end result is a machine learning model that runs on Google’s servers, accessible via an API. Users can reach out to that model via the Google cloud API and get predictions on new images. For example, both Disney and Urban Outfitters have tested AutoML to identify objects in their online stores so users can search and filter with more terms. Thus, you could search for “blue backpack” on Urban Outfitters and see all the blue backpacks, even if the items were not tagged that way in the system.
Google’s Cloud AutoML is currently limited to images, and it’s is alpha. You need to apply for access to the alpha version, and there’s no guarantee you’ll get in right now. The vision part of AutoML is just the first part of several features planned for the product. Google did not mention cost, but it’s likely businesses will have to pay for API access to the models they create in AutoML.
Continue reading
How to Install Windows 10 in a Virtual Machine
If you have to deal with files you can't trust, need to test multiple OS installations on the same system, or otherwise need access to the OS without wanting to use it as a daily driver, here's how to install Windows 10 in a virtual machine.
Windows 11 Will Launch on Oct. 5, but Only for Select Machines
Microsoft says Windows 11 will start hitting compatible PCs on October 5th. However, you might have to wait much, much longer to install it on your machine.
Microsoft Is Pushing the PC Health Check App to Windows 10 Machines
Don't want it? Tough, you're getting the Microsoft PC Health Check app regardless of whether or not you've expressed interest in Windows 11.
Windows Machines Need Up to Eight Hours to Update Successfully
Microsoft has found that devices with less than eight hours of Update Connectivity have a much higher chance of falling behind on updates.