Talk:1881: Drone Training

Deep learning

With deep learning (or any kind of actively learning Artificial Intelligence), drones can learn from their experience. The question is if you need spray bottle for this (that is, if there is some programmatic way to teach it, like marking parts of home that it should not enter in an app, or a feedback from app), and if the spraying the drone would actually work (if the AI would actually take it for a negative reinforcement). --JakubNarebski (talk) 10:33, 25 August 2017 (UTC)

Imho it would be really cool if you could teach an AI things without using any software, e.g. no programmatic way or app needed. Personally I'd prefer using a spray bottle over an app or similar. And I'm saying this as a professional software developer. Elektrizikekswerk (talk) 10:55, 25 August 2017 (UTC)

60 years ago when televisions had vacuum tubes instead of integrated circuits and the picture would get fuzzy, you could teach it to behave by slapping the side of the case. *wink* Rtanenbaum (talk) 12:02, 25 August 2017 (UTC)

Indeed, but it didn't learn from it XD /edit: I'm not sure if you're exaggerating with the "60 years ago", though... It's more like 10 to 15 when CRTs vanished https://en.wikipedia.org/wiki/Cathode_ray_tube#Demise. Elektrizikekswerk (talk) 13:01, 25 August 2017 (UTC)

"This is the first comic to feature Black Hat, White Hat, and no other character." Is this correct. Which other comics feature both "hats".Zeimusu (talk) 14:15, 25 August 2017 (UTC)

I've removed this from the explanation. Even when this is true it's not worth to be mentioned unless this really has a connection to the content of the comic. --Dgbrt (talk) 15:24, 25 August 2017 (UTC)

Is that interpretation of the alt text likely? My initial thought was that Black Hat's Roomba would be taken custody of because of the Prevention of Cruelty to Animals Act. (Well, actually, it was more like "aren't shock collars a form of animal abuse?") 162.158.2.166 16:43, 25 August 2017 (UTC)