Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
AI based home automation
#1
I ran into this when researching some Smartthings stuff. supposedly a cloud based AI service.

http://home.ai
|Z-Wave|Sonos|Tivo|Hue|Plex|Roku|MyMovies|Echo|
Nest|Harmony|Neeo|LG TV|Smarthings|
Reply
#2
(11-02-2017, 09:02 AM)potts.mike Wrote: I ran into this when researching some Smartthings stuff. supposedly a cloud based AI service.

http://home.ai

"The Home.AI platform gathers data from connected devices, smart phones, tracking tokens and external data sources. In real-time, deep learning algorithms detect patterns from these inputs, derives role and context from these patterns, and then automatically takes actions on behalf services can be delivered to family members based on location, context and observed behavior."

Can it ground my kids if they don't finish their homework?
_______________
Denon 3808ci, 2112ci , SONOS, Russound CAA66,Z-Wave(Lights,Locks), Hue, SmartThings,
iPads,Tivo,Hikvision,Elk-M1,iTunes,TED5000,Somfy RTS blinds, Amazon Echo, Google Home, Polk XRT12,
Honeywell Wi-Fi 9000 & Caleo Wi-Fi Thermostats, Rainmachine
Reply
#3
Dean- can you add this to the list please :-). AI like this is where HA is going to be heading - its the natural progression for it.
Mykel Koblenz
Illawarra Smart Home
Reply
#4
If they would let you partner with them it could be a source of recurring revenue.
|Z-Wave|Sonos|Tivo|Hue|Plex|Roku|MyMovies|Echo|
Nest|Harmony|Neeo|LG TV|Smarthings|
Reply
#5
The thing is, I don't think it's worth it. I mean, think about how it could possibly work. If it just watches what you do and then decides to do things on its own, that's guaranteed to not be good. You would almost certainly have to tell it what kinds of things you want it to pay attention to, what things are off limits for it, what to do in response to particular patterns, etc... If that's the case, I could do better than that myself and we could avoid the cloud stuff.

I would want to see it used in the real world and to be considered even remotely valuable before I'd commit to it, because I think that, without a bunch of user input, it's worthless. And, if you have to provide a bunch of user input, then you can programmatically do those things much more efficiently and all locally, and within the context of CQC where it access to everything (which that device wouldn't have.)

I've been continuing to study up on neural networks, but I just don't see this as a good application for it. It doesn't have enough access to information, or the ability to understand what the information it might be given means unless you spend a lot of time telling it what it means. If it were build into CQC, you wouldn't have to do that. It knows what things are.
Dean Roddey
Software Geek Extraordinaire
Reply
#6
A couple things

1. It doesn't have to do things on its own it can make suggestions that the user can choose to implement or not.
2. We're not trying to do some super high level analytics. It's lights on, lights off etc... The sample size does not have to be that big to detect these patterns
3.You are working on the assumption that all your users know all the capabilities of CQC. Being a user for several years I don't, its just to complicated for someone that is not fully immersed in the IT world. Surfacing things to "simple" users can only help to engage users.
|Z-Wave|Sonos|Tivo|Hue|Plex|Roku|MyMovies|Echo|
Nest|Harmony|Neeo|LG TV|Smarthings|
Reply
#7
I'm not saying the idea is bad. But the thing is, why have something learn that you want to turn the kitchen lights on at 8AM on weekdays, when you can just tell CQC to turn the kitchen lights on at 8AM on weekdays and not have to worry about whether it is going to do it or not or wait weeks for it to decide this is something you want to do?

OTOH, I definitely see the power of an *interactive artificial* intelligence, that you talk to and ask to do things and ask it to do things at particular times and stuff like that. I just don't get the benefits of the extra complexity of an external thing that has to learn what to do, which might take a long time, and when those things are quite simple to the point that it's sort of the equivalent of some scheduled events.

Making it easier to do some of those types of things could be pretty easily added to the auto-config system, to allow people to have us generate some common event scenarios for them.
Dean Roddey
Software Geek Extraordinaire
Reply


Possibly Related Threads...
Thread Author Replies Views Last Post
  Artificial intelligence for the home. potts.mike 6 1,104 08-29-2017, 01:55 PM
Last Post: potts.mike
  Google Home $99 batwater 27 6,354 05-18-2017, 09:24 AM
Last Post: IVB
  Get Started using AutoVoice on Google Home and Amazon Echo batwater 2 2,096 02-20-2017, 01:18 PM
Last Post: batwater
  Google Home batwater 4 1,019 10-05-2016, 06:38 PM
Last Post: Dean Roddey
  The Trackr - Room based occupation sensing possability znelbok 3 858 04-19-2016, 08:36 AM
Last Post: Dean Roddey
  Need Feedback on Automation Hubs MikeW 3 690 04-17-2016, 06:11 PM
Last Post: Dean Roddey
  Xandem Home jkmonroe 4 1,516 11-26-2015, 02:26 PM
Last Post: jkmonroe
  Automation Circa 1985 Dean Roddey 1 880 08-26-2015, 07:45 PM
Last Post: pinballmark
  slightly different perspective on Home automation from mouser SomeWhatLost 2 1,387 07-08-2015, 07:52 PM
Last Post: SomeWhatLost
  Nexia Home? Sendero 1 1,015 03-10-2015, 11:46 AM
Last Post: Dean Roddey

Forum Jump:


Users browsing this thread: 1 Guest(s)