Microsoft is trying to avoid updating Windows 10 while using PC using machine learning

by bruce mars

Periodic updating of the OS is necessary to repair OS vulnerabilities and add new functions. However, installation of Windows 10 is started while the device is in use, so restart may suddenly start while the user is using the PC. In order to avoid such a situation in Microsoft, it was reported that "We are trying not to update Windows 10 while using PC using machine learning".

Windows 10 now uses machine learning to stop updates installing when a PC is in use - The Verge
https://www.theverge.com/2018/7/25/17614842/microsoft-windows-10-updates-reboot-pc-machine-learning-feature

Windows10's overly aggressive update system, which sometimes restarts while using the device, has been a point of frustration for many users for a long time. Software has been developed to prevent forced update of Windows 10 by volunteers, but of course Microsoft also receives feedback from users and recognizes that there is a problem with Windows 10 update. Microsoft currently has a snooze feature to postpone installing updates, but it still seems that complaints of "restarting while using the device" are still being sent.

However, although you are dissatisfied with the OS update, you can not stop the update itself. So Microsoft announced that it is trying to relieve the user's pain associated with updating Windows 10 using machine learning. Train the machine learning model and predict the timing of restarting at the timing when it is on hold while updating it so that it is not most troubling for the user and restart the device at the optimum timing I said that it was.

The machine learning model says, "Now that the device is not in use, is it because the user has finished using the device, or is it just temporarily taking a break and planning to come back again?" Consider and make the best prediction, Microsoft says.

by Mike Licht

The machine learning model has already been tested inside Microsoft, and certain results have come out. If a machine learning model is introduced that avoids forced updates at annoying timing, great pain may be alleviated from users all over the world.