Overlap of learning and manual labeling?

Just installed yesterday, so please let me know if this is answered elsewhere or just over time through usage.

I know Sense is supposed to ‘learn’ which devices are represented in my power usage pattern, but I also just went through the settings to add all the devices in my home including make/model. So is Sense going to automatically integrate those two data sources?

I’m also happy to go around the house turning things on/off one by one if the app would allow me to manually input what device I’m toggling, but I don’t see that option. I recognize most users are not keen to do that, but I’m willing to! So how can I do that?

thanks!

… aaand now that I looked at the Knowledge Base as recommended for new users I found the “Can I help Sense learn my home?” topic and see that the answer is basically NO. I find that a little disappointing but am still excited to see what Sense learns (and to keep suggesting beta features…).

Here’s another useful read: Why can't you train Sense?

In short, we’d love to build such a feature, but it’s not a near-term option for the reasons noted above.

@danny.benjamin,
If you could mark the on-signature and the off-signature accurately within a 1/10 of a second, that might be useful for training, but most people can’t consistently get anywhere near that. Plus many appliances don’t reveal their real on and off signatures when switching on and off (refrigerators for one).

1 Like

You bring up an interesting point, which is that humans can generally not achieve the timing accuracy you speak of, but smart home devices could. I see a lot of other threads on here about smart home devices so I’ll leave that be, but I can’t help but wonder if there are opportunities there for added pathways (that will become more prevalent in the future) to identifying devices beyond the most low level energy draw signature.

@danny.benjamin,
Sense has already begun to exploit smart home technologies, as you suggest, with some of their integrations. The integrated TP-Link and Wemo smarplugs are gathering data for that very purpose - much better training for devices that Sense is having trouble with. And I suspect the Hue integration and the NDI detection are doing the same. With likely more to come.

@kevin1 has already noted smart plugs (learn more here). There’s also our Philips Hue integration. But there’s definitely more possibilities there for further integration with other smart products.