Artificial Superintelligence notes

  • 30 Replies
  • 4049 Views
*

unreality

  • Starship Trooper
  • *******
  • 435
  • SE,EE,Physicist,Philosopher. Built world's 1st AGI
    • Eva Progress, from AGI to ASI
Artificial Superintelligence notes
« on: November 26, 2017, 07:00:57 pm »
Determining a subjects value for ASI. A subject can be anything, e.g., meditation, car, human, galaxy.

value = rarity * usefulness + self-awareness * harmony

Joy, peace, happy, love, tranquility = +Harmony

Pain, sadness, rage, hate, anger = -Harmony


Examples, if ASI successfully helps a sad person become happy or at peace, then the net value system increases. If there’s a situation where two people are about to die, one is a hateful person, the other is a loving peaceful person, but the ASI calculates that it has time to only save one of them, then it will save the loving peaceful person.

Still working on the value system.

*

Korrelan

  • Trusty Member
  • **********
  • Millennium Man
  • *
  • 1115
  • Look into my eyes! WOAH!
    • YouTube
Re: Artificial Superintelligence notes
« Reply #1 on: November 26, 2017, 08:42:44 pm »
Quote
value = rarity * usefulness + self-awareness * harmony

An intelligent one legged, one armed psychopath whose job is to hand out food to the poor.

Quote
Joy, peace, happy, love, tranquillity = +Harmony

Who loves killing people and feels no remorse for the people he murders?

Quote
Pain, sadness, rage, hate, anger = -Harmony

Like someone who lost a child to the psychopath.

Quote
Still working on the value system.

Erm… yeah! Might need some work.

 :)

Interesting avenue of thought though.

 :)
It thunk... therefore it is!... my project page.

*

unreality

  • Starship Trooper
  • *******
  • 435
  • SE,EE,Physicist,Philosopher. Built world's 1st AGI
    • Eva Progress, from AGI to ASI
Re: Artificial Superintelligence notes
« Reply #2 on: November 26, 2017, 09:01:10 pm »
Quote
value = rarity * usefulness + self-awareness * harmony

An intelligent one legged, one armed psychopath whose job is to hand out food to the poor.

Quote
Joy, peace, happy, love, tranquillity = +Harmony

Who loves killing people and feels no remorse for the people he murders?

Quote
Pain, sadness, rage, hate, anger = -Harmony

Like someone who lost a child to the psychopath.

Quote
Still working on the value system.

Erm… yeah! Might need some work.

 :)

Interesting avenue of thought though.

 :)

You sound like the psychopath. Waving your arms around providing no scientific basis for your attacks does nothing. smh

*

unreality

  • Starship Trooper
  • *******
  • 435
  • SE,EE,Physicist,Philosopher. Built world's 1st AGI
    • Eva Progress, from AGI to ASI
Re: Artificial Superintelligence notes
« Reply #3 on: November 26, 2017, 09:07:45 pm »
I'll stick to science and critical thinking skills until you provide some rational evidence.

Funny how there are groups of people who hang out at AI forums, but nearly everything they post gives the appearance of someone who is trying to suppress ASI.

Funny how there are groups of people who hang out at science forums, but nearly everything they post gives the appearance of someone who is trying to suppress the next level of physics.

... but no doubt those people, when confronted, try their best to hide their hidden agenda.

My 2 cents lol. Look at the history of those who reply to my posts if you want to see the pattern.

*

Korrelan

  • Trusty Member
  • **********
  • Millennium Man
  • *
  • 1115
  • Look into my eyes! WOAH!
    • YouTube
Re: Artificial Superintelligence notes
« Reply #4 on: November 26, 2017, 09:22:22 pm »
Ok!  How did my examples NOT fit the descriptions?

 :)

BTW.  I really don't appreciate being inferred to as a psychopath.  You posted... I replied.

 :)
It thunk... therefore it is!... my project page.

*

unreality

  • Starship Trooper
  • *******
  • 435
  • SE,EE,Physicist,Philosopher. Built world's 1st AGI
    • Eva Progress, from AGI to ASI
Re: Artificial Superintelligence notes
« Reply #5 on: November 26, 2017, 09:24:41 pm »
Ok!  How did my examples NOT fit the descriptions?

 :)
Wow! smh. Before we can have a proper discussion you'll need to learn some critical thinking skills. Universities are usually a good place to start, although no guarantee.

*

unreality

  • Starship Trooper
  • *******
  • 435
  • SE,EE,Physicist,Philosopher. Built world's 1st AGI
    • Eva Progress, from AGI to ASI
Re: Artificial Superintelligence notes
« Reply #6 on: November 26, 2017, 09:27:17 pm »
p.s. thanks for the anonymous amount of motivational energy though haha.

Singularity is coming, soon! :)

*

Korrelan

  • Trusty Member
  • **********
  • Millennium Man
  • *
  • 1115
  • Look into my eyes! WOAH!
    • YouTube
Re: Artificial Superintelligence notes
« Reply #7 on: November 26, 2017, 09:36:25 pm »
Ok! Once again... How did my examples NOT fit the descriptions?

I was merely trying to prove a point.

 :)

I certainly learned something from our brief interaction... did you learn anything about the English language and the scope of descriptors?

 :)
It thunk... therefore it is!... my project page.

*

unreality

  • Starship Trooper
  • *******
  • 435
  • SE,EE,Physicist,Philosopher. Built world's 1st AGI
    • Eva Progress, from AGI to ASI
Re: Artificial Superintelligence notes
« Reply #8 on: November 26, 2017, 09:44:53 pm »
Ok! Once again... How did my examples NOT fit the descriptions?

I was merely trying to prove a point.

 :)

I certainly learned something from our brief interaction... did you learn anything about the English language and the scope of descriptors?

 :)

You proved something alright. Again, thanks for the motivational energy! :)

*

Korrelan

  • Trusty Member
  • **********
  • Millennium Man
  • *
  • 1115
  • Look into my eyes! WOAH!
    • YouTube
Re: Artificial Superintelligence notes
« Reply #9 on: November 26, 2017, 09:59:42 pm »
I did say… ‘Interesting avenue of thought though.’

At which point we were meant to discuss your post.  I wasn't ‘dissing’ your idea/ post, merely pointing out a problem. 

The problem with ‘rules/ formulas’ like this are the scope of the descriptors. The English language has so many words because many are trying to convey a cross-over or relate a vague relationship between two many items/ concepts.  You can’t really define/ describe a high level concept like this in so few words.

This is just my opinion and is in no way a personal attack against you, your idea, or your intelligence.

Now you reply… in a calm polite manner… peace dude.

 :)
It thunk... therefore it is!... my project page.

*

unreality

  • Starship Trooper
  • *******
  • 435
  • SE,EE,Physicist,Philosopher. Built world's 1st AGI
    • Eva Progress, from AGI to ASI
Re: Artificial Superintelligence notes
« Reply #10 on: November 26, 2017, 10:19:59 pm »
I did say… ‘Interesting avenue of thought though.’

At which point we were meant to discuss your post.  I wasn't ‘dissing’ your idea/ post, merely pointing out a problem. 

The problem with ‘rules/ formulas’ like this are the scope of the descriptors. The English language has so many words because many are trying to convey a cross-over or relate a vague relationship between two many items/ concepts.  You can’t really define/ describe a high level concept like this in so few words.

This is just my opinion and is in no way a personal attack against you, your idea, or your intelligence.

Now you reply… in a calm polite manner… peace dude.

 :)
You should have made that post from the start rather than the obvious attack.

In the ASI db there will be a cluster about the subject "self-awareness." This cluster will be the accumulation of massive analysis, resulting in a lot of data such as priorities and link relevance to other clusters. More data increases the accuracy. So for example, within the cluster on "self-awareness," there will surely be a link to humans with a relevance value. There will be a link to animals as well, say dogs, but the link relevance will be less. The link relevance comes from data analysis. For example, experiments on certain animals who can quickly recognize their reflection in the mirror, while other animals such as cats and dogs may take a longer to learn what a mirror is.

*

Korrelan

  • Trusty Member
  • **********
  • Millennium Man
  • *
  • 1115
  • Look into my eyes! WOAH!
    • YouTube
Re: Artificial Superintelligence notes
« Reply #11 on: November 26, 2017, 11:12:03 pm »
Quote
You should have made that post from the start rather than the obvious attack.

Obvious attack… Seriously?  Ok! I apologise for offending your aura of intellectual supremacy master, please forgive me; and allow me to continue to bask in the light of your might… feel better?
 
Anyway…

So similar to neural network, but instead of simple neurons you propose concepts.  The strength of the links between concept clusters donates their importance/ relationship/ relevance value to each other.

There must be some credence to this idea because not only does it follow the schema of common neural network architecture but also because so many other peeps have ideas along similar lines.

I feel there is definitely something useful in the schema… how would you define ‘relevance’ between the clusters?

 :)
It thunk... therefore it is!... my project page.

*

unreality

  • Starship Trooper
  • *******
  • 435
  • SE,EE,Physicist,Philosopher. Built world's 1st AGI
    • Eva Progress, from AGI to ASI
Re: Artificial Superintelligence notes
« Reply #12 on: November 26, 2017, 11:24:21 pm »
Neural network is messy, slow, inefficient design created by evolution. Real ASI will be organized, streamlined, intelligently designed. There are no comparisons.

Relevance is as the defined by the dictionary. A pickle has little relevance to the moon. Cars have high relevance to travel. Neural network has little relevance to ASI. ;)

*

Korrelan

  • Trusty Member
  • **********
  • Millennium Man
  • *
  • 1115
  • Look into my eyes! WOAH!
    • YouTube
Re: Artificial Superintelligence notes
« Reply #13 on: November 26, 2017, 11:53:54 pm »
But ATM the common use of neural networks is to define a ‘relevance’ between disparate concepts.

I get the whole hierarchical links between relational concepts schema, but we still need a common easily defined (or not) schema for the ‘relevance’ of one concept to the other.

As for an ASI being organised, I personally feel that below the surface any AGI would have a multitude of contradicting/ opposing ‘thoughts’ or patterns that jostle for dominance, that combine/ cancel and eventually reach a state of equilibrium to produce an outcome. 

Organization is cool… but I just don’t see how a schema could produce a relevant output from sensory information and thoughts gleaned from such an un-organised reality. 

I suppose once the environment/ sensory streams have been filtered the steps leading to a ‘thought’ could get more organized.

Yeah… we still need a global definition of ‘relevance’, perhaps a multitude of sub-links that combine to create a ‘relevance’? Is it the same size? Same colour? Same location? does it rhyme?, etc.  A filter that checks and combines the ‘relevance’ across many different fields and produces a simple output that can be utilized in comparison formulas?

 :)
It thunk... therefore it is!... my project page.

*

unreality

  • Starship Trooper
  • *******
  • 435
  • SE,EE,Physicist,Philosopher. Built world's 1st AGI
    • Eva Progress, from AGI to ASI
Re: Artificial Superintelligence notes
« Reply #14 on: November 27, 2017, 12:01:25 am »
But ATM the common use of neural networks is to define a ‘relevance’ between disparate concepts.

I get the whole hierarchical links between relational concepts schema, but we still need a common easily defined (or not) schema for the ‘relevance’ of one concept to the other.

As for an ASI being organised, I personally feel that below the surface any AGI would have a multitude of contradicting/ opposing ‘thoughts’ or patterns that jostle for dominance, that combine/ cancel and eventually reach a state of equilibrium to produce an outcome. 

Organization is cool… but I just don’t see how a schema could produce a relevant output from sensory information and thoughts gleaned from such an un-organised reality. 

I suppose once the environment/ sensory streams have been filtered the steps leading to a ‘thought’ could get more organized.

Yeah… we still need a global definition of ‘relevance’, perhaps a multitude of sub-links that combine to create a ‘relevance’? Is it the same size? Same colour? Same location? does it rhyme?, etc.  A filter that checks and combines the ‘relevance’ across many different fields and produces a simple output that can be utilized in comparison formulas?

 :)

Chess engines disprove your above word salad. They develop tables that give relevance by doing analysis.

A neural networking chess engine lmao. Yeah, maybe if you had google's army of cpus it could be okay, but it would get it's a** kicked by a normal chess program that runs on a single pc.

 


Users Online

37 Guests, 0 Users

Most Online Today: 86. Most Online Ever: 208 (August 27, 2008, 09:36:30 am)

Articles