Re: [agi] Who are you?

[Costi Dumitrescu]

Turing's is flawed

https://media.wired.com/photos/59323151edfced5820d0edfe/master/w_616,c_limit/recaptcha_anc...@2x.gif

On 27.03.2019 14:15, A.T. Murray wrote:

Sometimes a great notion occurs to us as we code the first working
artificial general intelligence (AGI). Yesterday was such a time. For
several years we had racked our brain with the problem of how the AI
Mind would be able to respond properly and separately to the questions
"Who are you?" and "What are you?". When the tentative solution
announced itself yesterday, we went to fetch pen and paper and we
wrote the idea down. We were about to code the solution in Perl or
Forth, but we got sidetracked by a TV movie and so now we tell the
AGI-list about the solution before we even code it into the first
working AGI software.

We were frustrated in recent years by not knowing how to get the AI
Mind to distinguish between "who" and "what" as interrogative
pronouns. The solution of yesterday is to attack the problem not
head-on but along indirect pathways.

First we plan to have the http://ai.neocities.org/InStantiate.html
module trap the incoming pronoun "who" by causing activation to be
imposed upon such who-related concepts as "person" and "name".

Then we need to expand the http://ai.neocities.org/SpreadAct.html
module to make it spread activation not to just one target-concept at
a time, but to other concepts associated with the target-concept.
Typically an AI Mind may know the following facts.

Andru is a person.
My name is Andru.
I am Andru.
I am a robot.
I need a body.

If we ask the AI "Who are you?" we will then expect activations to
result.

"Who" activates "person" and "name" as concepts.
"Are" activates the concept of the verb "be".
"You" activates the self-concept of "I" or ego.

When the perhaps-not-conscious-but-somewhat-self-aware AI activates
ideas about "self" in the knowledge-base, we can expect the AI to
answer "I AM ANDRU" because of the convergence of activation along
indirect pathways upon stored ideas containing the word "Andru".

*Artificial General Intelligence List
* / AGI / see discussions
 + participants
 + delivery options
 Permalink




--
Artificial General Intelligence List: AGI
Permalink: 
https://agi.topicbox.com/groups/agi/T867c6b40569d579c-Ma3a125c70ce5de3c0d644133
Delivery options: https://agi.topicbox.com/groups/agi/subscription

[agi] Who are you?

[A.T. Murray]

Sometimes a great notion occurs to us as we code the first working
artificial general intelligence (AGI). Yesterday was such a time. For
several years we had racked our brain with the problem of how the AI Mind
would be able to respond properly and separately to the questions "Who are
you?" and "What are you?". When the tentative solution announced itself
yesterday, we went to fetch pen and paper and we wrote the idea down. We
were about to code the solution in Perl or Forth, but we got sidetracked by
a TV movie and so now we tell the AGI-list about the solution before we
even code it into the first working AGI software.

We were frustrated in recent years by not knowing how to get the AI Mind to
distinguish between "who" and "what" as interrogative pronouns. The
solution of yesterday is to attack the problem not head-on but along
indirect pathways.

First we plan to have the http://ai.neocities.org/InStantiate.html module
trap the incoming pronoun "who" by causing activation to be imposed upon
such who-related concepts as "person" and "name".

Then we need to expand the http://ai.neocities.org/SpreadAct.html module to
make it spread activation not to just one target-concept at a time, but to
other concepts associated with the target-concept. Typically an AI Mind may
know the following facts.

Andru is a person.
My name is Andru.
I am Andru.
I am a robot.
I need a body.

If we ask the AI "Who are you?" we will then expect activations to result.

"Who" activates "person" and "name" as concepts.
"Are" activates the concept of the verb "be".
"You" activates the self-concept of "I" or ego.

When the perhaps-not-conscious-but-somewhat-self-aware AI activates ideas
about "self" in the knowledge-base, we can expect the AI to answer "I AM
ANDRU" because of the convergence of activation along indirect pathways
upon stored ideas containing the word "Andru".

--
Artificial General Intelligence List: AGI
Permalink: 
https://agi.topicbox.com/groups/agi/T867c6b40569d579c-M178fb2a15df45c8eee9009ac
Delivery options: https://agi.topicbox.com/groups/agi/subscription