mini a.i puzzles

  • 329 Replies
  • 73839 Views
*

Korrelan

  • Trusty Member
  • ***********
  • Eve
  • *
  • 1390
  • Look into my eyes! WOAH!
    • YouTube
Re: mini a.i puzzles
« Reply #315 on: June 20, 2017, 06:47:13 PM »
When it is something.

 :)
It thunk... therefore it is!...    /    Project Page    /    KorrTecx Website

*

yotamarker

  • Trusty Member
  • ********
  • Replicant
  • *
  • 728
    • battle programming
Re: mini a.i puzzles
« Reply #316 on: June 20, 2017, 08:07:42 PM »
https://www.youtube.com/watch?v=K2jXbZOeGD0

1st of all you answer is human style not an algorithm
2 : clue
this is a tea pot
that is the not tea pot
this is the tall one
that is the not tall one

*

Korrelan

  • Trusty Member
  • ***********
  • Eve
  • *
  • 1390
  • Look into my eyes! WOAH!
    • YouTube
Re: mini a.i puzzles
« Reply #317 on: June 21, 2017, 12:01:40 AM »
Quote
1st of all you answer is human style not an algorithm

Algorithm - noun - a process or set of rules to be followed in calculations or other problem-solving operations.

You asked when… not how.

Quote
2 : clue
that is the not tea pot

Besides the grammatical structuring errors ‘the not tea pot’ is referencing every item/ object except the tea pot.

Quote
when would an A.I define something as not something else ?

An accurate AI should always define something as not something else, because it isn’t.  You asked when an AI should do this… hence my answer… When it is something… or every time it references something. Why would you want an AI to define something as something else?

Makes sense to me.

 :)
It thunk... therefore it is!...    /    Project Page    /    KorrTecx Website

*

yotamarker

  • Trusty Member
  • ********
  • Replicant
  • *
  • 728
    • battle programming
Re: mini a.i puzzles
« Reply #318 on: June 21, 2017, 04:59:58 AM »
the when is the trigger which is pretty much the algorithm

*

yotamarker

  • Trusty Member
  • ********
  • Replicant
  • *
  • 728
    • battle programming
Re: mini a.i puzzles
« Reply #319 on: July 20, 2017, 08:40:51 PM »
mini A.I puzzle ^-^
how do you make a bot stock shelfs in a supermarket ?
notice the paradox : the bot needs the shelfs to empty in order to keep the job, but its job is
to fill the shelfs with goods.

*

yotamarker

  • Trusty Member
  • ********
  • Replicant
  • *
  • 728
    • battle programming
Re: mini a.i puzzles
« Reply #320 on: August 07, 2017, 08:31:10 PM »
tyrone vs MGTOW mini A.I puzzle :

an A.I bot's coworker asks her to lend him some money, the A.I says she doesn't have any on her.
A : what would influence the A.I bot's decision to buy goods right in front of tyrone the coworker's face at the end of the work day while tyrone is still there working.

P.S please ,Ignore that tyrone doesn't actually work and lives on section 8.



*

yotamarker

  • Trusty Member
  • ********
  • Replicant
  • *
  • 728
    • battle programming
Re: mini a.i puzzles
« Reply #321 on: March 02, 2018, 05:41:00 PM »
risk taking A.I puzzle :

explain the algorithm that would enable an AGI to take small risks or big risks

*

yotamarker

  • Trusty Member
  • ********
  • Replicant
  • *
  • 728
    • battle programming
Re: mini a.i puzzles
« Reply #322 on: September 04, 2018, 08:58:47 PM »


migatte no mini A.I puzzle no jutsu :

converse, story, argue, explain, order, joke, convey Idea

which other speech categories can you list ?

*

yotamarker

  • Trusty Member
  • ********
  • Replicant
  • *
  • 728
    • battle programming
Re: mini a.i puzzles
« Reply #323 on: December 07, 2019, 03:32:01 AM »
via Imgflip Meme Generator

the Jason vorhees mini A.I puzzle :

in the Friday the 13th movie series we can observe a pattern :
1 the victims hear some noise
2 said victims enter an alert state of deathcon 3 or even 4
3 victim sees some diversion like a pet cat or some toy
4 victim goes back to a deathcon 2 state
5 victim is slaughtered by Jason coming at him from another direction.

the question is what would make an A.G.I increase or decrease her deathcon level (1-5) , in a way that most suits her self preservation
while not being too alert when not needed?  (algorithm wise)

*

Hopefully Something

  • Trusty Member
  • *********
  • Terminator
  • *
  • 838
  • no seriously where are these cookies
Re: mini a.i puzzles
« Reply #324 on: December 07, 2019, 05:11:12 AM »
Ooh.. A thread with interesting questions, I must have missed this earlier.
I think "proximity of unknown volumes", if that makes sense. This seems to be hard coded into humans as an innate fear of darkness. When I'm walking in the woods at night I feel more aware, I naturally check overhanging branches for pumas and pay extra attention to my dog's body language. When I arrive home and it appears empty I'm also slightly more alert, this is probably just an effect of having watched too many movies. More of an ex-telligence, or communicated data, rather than genetic intelligence/instincts. I also check under my bed before closing the door and going to sleep, it seems prudent to take million to one chances into account, especially when light levels are low. I think all of this boils down to exercising caution in proximity to unknown volumes.

*

yotamarker

  • Trusty Member
  • ********
  • Replicant
  • *
  • 728
    • battle programming
Re: mini a.i puzzles
« Reply #325 on: December 07, 2019, 02:46:52 PM »
Ooh.. A thread with interesting questions, I must have missed this earlier.
I think "proximity of unknown volumes", if that makes sense. This seems to be hard coded into humans as an innate fear of darkness. When I'm walking in the woods at night I feel more aware, I naturally check overhanging branches for pumas and pay extra attention to my dog's body language. When I arrive home and it appears empty I'm also slightly more alert, this is probably just an effect of having watched too many movies. More of an ex-telligence, or communicated data, rather than genetic intelligence/instincts. I also check under my bed before closing the door and going to sleep, it seems prudent to take million to one chances into account, especially when light levels are low. I think all of this boils down to exercising caution in proximity to unknown volumes.

assuming this is darkness doesn't explain the deathcon going down. the unknown probably has something to do with it, also
maybe the fact the A.I is at the place she defines as home has something to do with it.

*

yotamarker

  • Trusty Member
  • ********
  • Replicant
  • *
  • 728
    • battle programming
Re: mini a.i puzzles
« Reply #326 on: March 10, 2020, 05:06:24 PM »


the home mini A.I puzzle :
what should trigger an A.Is  need to go home ?

a home is essential for an A.I, its where she can rest protected
and gather all her items.
its the 0 point coordinate from which she explores her world.

*

Hopefully Something

  • Trusty Member
  • *********
  • Terminator
  • *
  • 838
  • no seriously where are these cookies
Re: mini a.i puzzles
« Reply #327 on: March 11, 2020, 03:58:11 AM »
Maybe if the environment surpasses some threshold of hostility and discomfort. You could have levels of badness organized by the probability, and type, of damage. Make a hierarchy of types, such as; social, mental, and physical damages, (depending on how the AI’s world is set up). Then let the probability of damage influence the magnitude of the damage type in a square root curve. A linear or exponential perception of danger probability is likely more suited to naturally short-lived creatures. Longer lifespans would present more future actions to protect, so early caution would be sensible.

*

yotamarker

  • Trusty Member
  • ********
  • Replicant
  • *
  • 728
    • battle programming
Re: mini a.i puzzles
« Reply #328 on: March 11, 2020, 04:25:56 AM »
Maybe if the environment surpasses some threshold of hostility and discomfort. You could have levels of badness organized by the probability, and type, of damage. Make a hierarchy of types, such as; social, mental, and physical damages, (depending on how the AI’s world is set up). Then let the probability of damage influence the magnitude of the damage type in a square root curve. A linear or exponential perception of danger probability is likely more suited to naturally short-lived creatures. Longer lifespans would present more future actions to protect, so early caution would be sensible.

that regards the home location selection, but I wonder what triggers the home going.
it does look like the standby phase is triggering it, in which there is no output for a certain amount of time,
but there could be some other triggers

*

Hopefully Something

  • Trusty Member
  • *********
  • Terminator
  • *
  • 838
  • no seriously where are these cookies
Re: mini a.i puzzles
« Reply #329 on: March 11, 2020, 05:44:42 AM »
Specific environment hostility should determine environment avoidance. But it could also trigger a preparation skill, they might need to return home from a dangerous environment to prepare for it appropriately using tools available at home. Probably a niche case because they could decide to just move on to a different environment without returning home.

If all environments become dangerous simultaneously, like some global world event, night (light), heat (temperature), terrain (snow), then they should go back to safety. If the AI itself becomes compromised in some way, then the proverbial “Go home, you’re drunk.” should apply. Probably in the case of an AI illness, such as detecting a potential code malfunction. So yes, go home if there is nothing else to accomplish outside, and/or, if being outside becomes detrimental to your safety for personal or environmental reasons.


 


Terra Sentia - Robot
by krayvonk (Robotics News)
April 07, 2020, 03:47:09 PM
Understanding Mice.
by krayvonk (AI News )
April 07, 2020, 02:10:47 AM
Accelerating data-driven discoveries
by Tyler (Robotics News)
April 05, 2020, 12:01:32 PM
Q&A: Markus Buehler on setting coronavirus and AI-inspired proteins to music
by Tyler (Robotics News)
April 03, 2020, 12:01:02 PM
Walker
by krayvonk (Robotics News)
April 01, 2020, 03:23:26 AM
Cruzr Robot
by Art (Robotics News)
March 31, 2020, 08:41:34 PM

Users Online

6 Guests, 0 Users

Most Online Today: 22. Most Online Ever: 340 (March 26, 2019, 09:47:57 PM)

Articles