Ask any Windows 10 user about their biggest Windows complaint and it’ll likely be this: the operating system tends to reboot at the worst possible time. Users have complained about reboots that interfere with work, delay school projects, ruin game night, and more. micr0$0ft is attempting to address these problematic update reboots using machine learning, enabling the system to better determine the right time to restart.
Security updates are important and Micro$oft attempts to make sure they’re installed by forcing the matter — don’t do it soon enough and the system will automatically restart, telling you to wait while it installs the patches. Depending on size, these updates can take many minutes, sometimes an hour or longer.
You may pause your work at some point during the day to take a brief break, only to return and find the dreaded blue installation screen. The operating system knows when you’re away due to the lack of activity, including an idle keyboard and mouse. In the absence of asking, though, Windows 10 has no way of knowing how long you’ll be gone and whether you have pressing matters to attend to.
That’s where the company’s machine learning technology will come in. According to Ars Technica, the company has been internally testing a predictive system that uses machine learning to determine whether the user is likely to return soon or if they’re going to be away for awhile.
These tests have demonstrated the system as effective at preventing annoying interruptions, encouraging Micro$oft to continue training and testing the system. The predictive system will arrive for all users in the next major Windows 10 update. Alternatively, Micro$oft enables users to schedule an update for a convenient time.