Sense Blog
turning on a light

Help Make Sense Smarter – Device Is Not On

While it’s constantly collecting information to help our Data Science Team and drive our machine learning algorithms, we also count on users to help train Sense, through new features like “Device Is Not On.”

Sense can detect what’s going on in your home primarily through machine learning. It’s constantly collecting information that’s helping our Data Science Team. But we also count on users like you to help train Sense, and that’s why we’re rolling out new features like “Device Is Not On.”

Is This Thing On?

Sometimes Sense will show a device as on, but in reality it’s not. What’s happening is that Sense sees the device turn on, but misses when it switches off. This is one of the most common errors users report. Of course, these “false ons” will decrease as we collect more data, but you can help make Sense more accurate by using the Device Is Not On feature through the app. To do so, go to a device’s details page, select “Report a Problem” at the bottom, and then “Device is Not On” from the menu.

We Love Your Feedback

Every time you share your real-world feedback, we have another data point to fine tune our models. This crowdsourced learning helps Sense understand what’s going on in your home and across the entire Sense community. Each notification helps, so don’t hesitate to share your experience.

Log These Scenarios

Your Sense app says your dishwasher is on, but you’re standing right next to it, and you can see that it is not on. Just take a second to log that event as “Device is Not On.” Or, what if you turn on your toaster, but Sense sees this as your coffee pot? This is another scenario that you should label as “Device Is Not On.” So keep an eye on how Sense is identifying devices in your home, and help us make Sense more accurate.

Help Train Sense

“Device is Not On” is just one of the ways you’re helping to train Sense. Contribute even more to the Sense community by categorizing and renaming devices, adding make/model information, logging your “guesses”, and enabling Network Device Identification. All these other data sources help build out a robust database of device signatures that will help the Sense technology grow and improve!