[agi] Maybe it is a question of learning how to communicate with limited AI

[Jim Bromer via AGI]

I am doing almost no programming (other than some simple website
stuff) but I have started to wonder if my basic ideas about language
might be used as a basis for improving web search engines. So I
started with my usual, fairly sophisticated, ideas, but then I started
wondering how I could get them to work with a search engine like
Watson. If you are going to try to Interface with an existing search
engine you are going to get stuck with their limitations. And then I
realized the question was not whether I could get Watson to hold a
conversation in order to make more intelligent searches, but whether I
could learn how it interprets certain forms of statements and then use
that as a way of communicating with it and thereby making it look more
like it is communicating with me. In the end I might not get it to
learn - through conversation - to find the kind of information that I
might be interested in, but I might be able to learn something about
the way it "thinks" and then use this to make a more sophisticated
user interface. I am seriously thinking about pursuing this. I assume
that IBM is still letting developers access Watson online in some
form.
Jim Bromer

--
Artificial General Intelligence List: AGI
Permalink: 
https://agi.topicbox.com/groups/agi/T417db08f2846a615-Md08e696e232fad074b511da1
Delivery options: https://agi.topicbox.com/groups

[agi] The four things needed to solve AGI.

[Alan Grimes]

Nanograte Knowledge Technologies via AGI wrote:
> Alan
>
> I suppose yo already possess a computational framework to process the
> research you are proposing as being critical? In other words, I think
> if you do not have the intelligence architecture, none of the
> functionality would be agi effective.


I'm missing a number of things that limits my ability to do this
research myself.

-> a GPU programming framework that actually works on my computer,
preferably OpenCL for portability reasons...

-> a simulation environment... Microsoft Malmo seems pretty close to
what I want (bare minimum actually), I'll have to take a look at it
again to see if I can make it run the way I need it to.

-> Everything is in python these days, so I'll need a high quality
manual for the python language. I don't think this documentation
actually exists. In fairness, I haven't made a concerted effort to
search for it recently. on-line courses in the subject only cover the
whitespace-based nesting but basically ignores all other parts of syntax
such as data definition and expression syntax.

-- 
Please report bounces from this address to a...@numentics.com

Powers are not rights.


--
Artificial General Intelligence List: AGI
Permalink: 
https://agi.topicbox.com/groups/agi/T507c404b4595c71c-Me0f8479ffd96b8e5e57b6721
Delivery options: https://agi.topicbox.com/groups

[agi] What is needed to bootstrap AGI

[Alan Grimes]

Okay, lets consider the concept of the design space.

The design space for motorized vehicles consists of just about anything
that has enough wheels to be stable on the ground, has a structural
frame, has an energy source, and has an engine of some kind to apply
that energy source, in the form of torque, to one or more of the wheels.
We have seen many examples of motorized vehicles for more than two
centuries.  ( https://en.wikipedia.org/wiki/History_of_steam_road_vehicles )

These days just about any bozo can produce a basic vehicle using off the
shelf parts. The key there is that a massive amount of effort was
required to produce those off the shelf parts. Today, all of the parts
that are available were designed for purposes other than AGI, meaning
that it will require a massive amount of effort to adapt them to
purpose, compounded by the fact that we don't quite know what we are
building.

Nevertheless, we must suspect that there is a set of parameters that
define the set of workable AGI designs and that this space is fairly
large. We also assume that we have freedom within a basic design to swap
one thing for another and still end up with a working design just as you
can swap factory rims with pimp rims or race wheels or opt for some more
rubber to give you a more LX ride. Similarly you can swap a steam engine
for a gas engine for an electric motor without violating the basic
concept of the motor vehicle. All of them will move down the road.

The first thing that is required for an AGI is a system complex enough
to allow it to manifest it's general intelligence. That is either a
robot of a reasonable complexity, or a simulation of some kind with
parameters of sufficient complexity that it demands an agent be capable
of generating concepts on the fly. Here is a game that I've wasted far
far too much of my life on, it's similar to Minecraft but it lets you
build starships, here's one I remodeled:

https://steamcommunity.com/sharedfiles/filedetails/?id=929823165

The critical point here is that the game only provides a few dozen block
types. It is your job to assign them a meaning or purpose and to create
concepts such as wings, compartments, hulls, etc... I'm not 100% certain
this is sufficient, but it is certainly a necessary level of complexity.
This is not an easy problem. Google is using Atari games these days.
Building a more satisfactory simulation is probably a hundred million
dollar development project. =~(

( I am flat broke w/ no income).

The AI's avatar should be reasonably similar to human modalities, at
least not completely alien.


Secondly, AI is not a blank slate. It definitely isn't just a grey block
of empty computronium

Now I'm not exactly sure what level of complexity is necessary, or
whether it is possible to start with a minimal kernel of some kind and
then bootstrap complexity from that, but there almost certainly does
need to be a framework sufficient enough for the AGI to start exhibiting
basic behaviors almost immediately. In many cases, the brain seems to be
cobbling together a solution for a high level algorithm using neurons
(the only available hammer...), in those cases, you can get a 1,000x+
speedup immediately from just implementing the high level code.

Normally, we would say that we just want to get it done, for the first
generation AGI. In this case, however, a 1,000x speed up means getting
to the singularity ten years sooner, so carefully chosen early
optimizations are a big part of getting it done.


Okay, my original idea was to base everything on abstractions of a form
similar to   A -> { W, X, Y, Z } and then construct a mind out of
several billion of those. The advantage to that concept is that it can
use perfectly conventional memory management techniques. The other
approach is to allocate a matrix that's Big Enough (tm) for your deep
neural network and basically accept that it has a learning-limit when it
hits its maximum entropy. At that point you have two options, really do
make it big enough to achieve enough intelligence to design version 2.0,
or design an introspection algorithm that can optimize and re-allocate
neural matricies, and basically patch up some of the pathological
conditions neural systems can get into.

Now the big problem that people are fretting about is how to "transfer
knowledge between tasks". There was even a [poorly constructed] contest
on this problem last year.   Want to know the secret? DON'T!  =P The
brain does not transfer knowledge between tasks, it simply re-uses
entire neural sub-units in different combinations to produce different
behaviors, or to solve different problems. While a linear neural model
seems to be sufficient to re-produce the quirks seen in human
psychology, it is clear that a switching network also exists that
dynamically reconfigures the brain for each task it performs.
Replicating this functionality is probably the most important problem
right now internal to the AGI itself.

Anyway, I'm getting off 

[agi] What is needed to bootstrap AGI

[Alan Grimes]

Okay, lets consider the concept of the design space.

The design space for motorized vehicles consists of just about anything
that has enough wheels to be stable on the ground, has a structural
frame, has an energy source, and has an engine of some kind to apply
that energy source, in the form of torque, to one or more of the wheels.
We have seen many examples of motorized vehicles for more than two
centuries.  ( https://en.wikipedia.org/wiki/History_of_steam_road_vehicles )

These days just about any bozo can produce a basic vehicle using off the
shelf parts. The key there is that a massive amount of effort was
required to produce those off the shelf parts. Today, all of the parts
that are available were designed for purposes other than AGI, meaning
that it will require a massive amount of effort to adapt them to
purpose, compounded by the fact that we don't quite know what we are
building.

Nevertheless, we must suspect that there is a set of parameters that
define the set of workable AGI designs and that this space is fairly
large. We also assume that we have freedom within a basic design to swap
one thing for another and still end up with a working design just as you
can swap factory rims with pimp rims or race wheels or opt for some more
rubber to give you a more LX ride. Similarly you can swap a steam engine
for a gas engine for an electric motor without violating the basic
concept of the motor vehicle. All of them will move down the road.

The first thing that is required for an AGI is a system complex enough
to allow it to manifest it's general intelligence. That is either a
robot of a reasonable complexity, or a simulation of some kind with
parameters of sufficient complexity that it demands an agent be capable
of generating concepts on the fly. Here is a game that I've wasted far
far too much of my life on, it's similar to Minecraft but it lets you
build starships, here's one I remodeled:

https://steamcommunity.com/sharedfiles/filedetails/?id=929823165

The critical point here is that the game only provides a few dozen block
types. It is your job to assign them a meaning or purpose and to create
concepts such as wings, compartments, hulls, etc... I'm not 100% certain
this is sufficient, but it is certainly a necessary level of complexity.
This is not an easy problem. Google is using Atari games these days.
Building a more satisfactory simulation is probably a hundred million
dollar development project. =~(

( I am flat broke w/ no income).

The AI's avatar should be reasonably similar to human modalities, at
least not completely alien.


Secondly, AI is not a blank slate. It definitely isn't just a grey block
of empty computronium

Now I'm not exactly sure what level of complexity is necessary, or
whether it is possible to start with a minimal kernel of some kind and
then bootstrap complexity from that, but there almost certainly does
need to be a framework sufficient enough for the AGI to start exhibiting
basic behaviors almost immediately. In many cases, the brain seems to be
cobbling together a solution for a high level algorithm using neurons
(the only available hammer...), in those cases, you can get a 1,000x+
speedup immediately from just implementing the high level code.

Normally, we would say that we just want to get it done, for the first
generation AGI. In this case, however, a 1,000x speed up means getting
to the singularity ten years sooner, so carefully chosen early
optimizations are a big part of getting it done.


Okay, my original idea was to base everything on abstractions of a form
similar to   A -> { W, X, Y, Z } and then construct a mind out of
several billion of those. The advantage to that concept is that it can
use perfectly conventional memory management techniques. The other
approach is to allocate a matrix that's Big Enough (tm) for your deep
neural network and basically accept that it has a learning-limit when it
hits its maximum entropy. At that point you have two options, really do
make it big enough to achieve enough intelligence to design version 2.0,
or design an introspection algorithm that can optimize and re-allocate
neural matricies, and basically patch up some of the pathological
conditions neural systems can get into.

Now the big problem that people are fretting about is how to "transfer
knowledge between tasks". There was even a [poorly constructed] contest
on this problem last year.   Want to know the secret? DON'T!  =P The
brain does not transfer knowledge between tasks, it simply re-uses
entire neural sub-units in different combinations to produce different
behaviors, or to solve different problems. While a linear neural model
seems to be sufficient to re-produce the quirks seen in human
psychology, it is clear that a switching network also exists that
dynamically reconfigures the brain for each task it performs.
Replicating this functionality is probably the most important problem
right now internal to the AGI itself.

Anyway, I'm getting off 

Re: [agi] The four things needed to solve AGI.

[MP via AGI]

I was addressing you, Rob :)

Sent from ProtonMail Mobile

On Fri, Jun 8, 2018 at 5:56 AM, Nanograte Knowledge Technologies via AGI 
 wrote:

> MP, were you addressing your comment at me, or Alan? If, at me, I could start 
> you off with an IEEE-reviewed document. The exact method in the white paper 
> is one of the many components that has to be programmed.
>
> Please let me know.
>
> Rob
> ---
>
> From: MP via AGI 
> Sent: Friday, 08 June 2018 8:34 AM
> To: AGI
> Subject: Re: [agi] The four things needed to solve AGI.
>
> Alan, would you be interested in emailing me these models? I’m an experienced 
> coder and would love to work on it.
>
> mindpixel at proton mail dot com
>
> Sent from ProtonMail Mobile
>
> On Fri, Jun 8, 2018 at 1:31 AM, Nanograte Knowledge Technologies via AGI 
>  wrote:
>
>> Alan
>>
>> Thank you for your response. I am in no way trying to challenge your 
>> personal research, or view. My questions are genuine and AGI related only. I 
>> do not agree with the view of some here who defend A.T. Murray's position by 
>> saying he at least does something. I agree with your view that it is simply 
>> not enough to just make an effort, therefore AGI would result. The true-AGI 
>> problem is many-fold more complex than we may even imagine. But all efforts 
>> may bring forth new inspiration.
>>
>> When compared to aspects of AGI being produced in robotic form today, I have 
>> no idea why Murray et al are trying to reinvent any one of the wheels that 
>> are clearly, already turning. It behooves progress to continually refine the 
>> real questions to be addressed, as you have obviously done. Perhaps, what 
>> would emerge from those efforts would find practical use?However, for its 
>> lack of structured debate between simple and complex systems, I would not be 
>> able to relate to it. That's my definitive view on Murray's world.
>>
>> I do not program anymore, but I have been designing computational theory for 
>> system of systems - to purpose - for a long time. AGI, as a holistic 
>> architecture, can be nothing less than a galaxy of deliberate system of 
>> systems, administered by at least 1, superpositioned system of systems. I 
>> have field tested my theoretical designs against real projects and with real 
>> people. The results have been accurate and reliable. The mature models are 
>> ready for programming.
>>
>> This is the intelligence (AGI) architecture I am referring to. In my view, 
>> it represents as a fully-recursive, adaptive, intelligence architecture (not 
>> intelligence bootstrapped), which would go a long way towards enabling AGI 
>> functionality, and beyond. Given the appropriate building blocks being put 
>> into place to operate within an intelligence architecture (world), the 
>> functionality you seek would in theory be able to evolve "naturally". Once 
>> programmed into a functional, matured, intelligent system, this may have 
>> potential to serve as a future, AGI hub, or docking station. Such is my 
>> theory.
>>
>> According to my analysis, a number of the key components of this 
>> intelligence architecture have already been programmed by other parties, 
>> eg., by CISCO, IBM, MS, GOOGLE, OPENCOG, and so many others in the big-wide 
>> world. Such progress serves to inform me that the overall architecture - 
>> formulating in my mind - still is on the right track. My passion continues 
>> as a hobby. I do it for the fun of it. I'm enjoying the journey. My purpose 
>> is not to compete with the efforts of others, but to indulge my life-long 
>> fascination and curiosity with the technological challenges presented by 
>> AGI, and beyond.
>>
>> In summary, the one thing needed to solve AGI is, in my opinion, still 
>> lacking in the public domain.
>>
>> Rob
>> ---
>>
>> From: Alan Grimes 
>> Sent: Friday, 08 June 2018 2:55 AM
>> To: Nanograte Knowledge Technologies via AGI
>> Subject: Re: [agi] The four things needed to solve AGI.
>>
>> Nanograte Knowledge Technologies via AGI wrote:
>>> Alan
>>>
>>> I suppose yo already possess a computational framework to process the
>>> research you are proposing as being critical? In other words, I think
>>> if you do not have the intelligence architecture, none of the
>>> functionality would be agi effective.
>> 
>> I'm missing a number of things that limits my ability to do this
>> research myself.
>> 
>> -> a GPU programming framework that actually works on my computer,
>> preferably OpenCL for portability reasons...
>> 
>> -> a simulation environment... Microsoft Malmo seems pretty close to
>> what I want (bare minimum actually), I'll have to take a look at it
>> again to see if I can make it run the way I need it to.
>> 
>> -> Everything is in python these days, so I'll need a high quality
>> manual for the python language. I don't think this documentation
>> actually exists. In fairness, I 

Re: [agi] The four things needed to solve AGI.

[johnrose]

>Alan
>
>We really really don't have any more time to waste on...
>

Agreed.  Compadres, we should not do FUD, gaslighting, trolling, etc..

>
>MP
>  BUT AT LEAST HE HAS SOMETHING.

As Google knows, searching is SOMETHING, but better to understand what one is 
searching for though often what we are searching for isn’t that which is what 
we are really after but just pieces in a complex systems mosaic of identity. 
IOW searching/researching is a dimension of creating/re-creating... us.

"We tirelessly and ceaselessly search for Something, we know not what, which 
will appear in the end to those who have penetrated to the very heart of 
reality."

--- Pierre Teilhard de Chardin

John

--
Artificial General Intelligence List: AGI
Permalink: 
https://agi.topicbox.com/groups/agi/T507c404b4595c71c-M21ddd2ee03417c1934e23ee5
Delivery options: https://agi.topicbox.com/groups

Re: [agi] The four things needed to solve AGI.

[Nanograte Knowledge Technologies via AGI]

MP, were you addressing your comment at me, or Alan? If, at me, I could start 
you off with an IEEE-reviewed document. The exact method in the white paper is 
one of the many components that has to be programmed.

Please let me know.

Rob

From: MP via AGI 
Sent: Friday, 08 June 2018 8:34 AM
To: AGI
Subject: Re: [agi] The four things needed to solve AGI.

Alan, would you be interested in emailing me these models? I’m an experienced 
coder and would love to work on it.

mindpixel at proton mail dot com

Sent from ProtonMail Mobile


On Fri, Jun 8, 2018 at 1:31 AM, Nanograte Knowledge Technologies via AGI 
mailto:agi@agi.topicbox.com>> wrote:
Alan

Thank you for your response. I am in no way trying to challenge your personal 
research, or view. My questions are genuine and AGI related only. I do not 
agree with the view of some here who defend A.T. Murray's position by saying he 
at least does something. I agree with your view that it is simply not enough to 
just make an effort, therefore AGI would result. The true-AGI problem is 
many-fold more complex than we may even imagine. But all efforts may bring 
forth new inspiration.

When compared to aspects of AGI being produced in robotic form today, I have no 
idea why Murray et al are trying to reinvent any one of the wheels that are 
clearly, already turning. It behooves progress to continually refine the real 
questions to be addressed, as you have obviously done. Perhaps, what would 
emerge from those efforts would find practical use?However, for its lack of 
structured debate between simple and complex systems, I would not be able to 
relate to it. That's my definitive view on Murray's world.

I do not program anymore, but I have been designing computational theory for 
system of systems - to purpose - for a long time. AGI, as a holistic 
architecture, can be nothing less than a galaxy of deliberate system of 
systems, administered by at least 1, superpositioned system of systems. I have 
field tested my theoretical designs against real projects and with real people. 
The results have been accurate and reliable. The mature models are ready for 
programming.

This is the intelligence (AGI) architecture I am referring to. In my view, it 
represents as a fully-recursive, adaptive, intelligence architecture (not 
intelligence bootstrapped), which would go a long way towards enabling AGI 
functionality, and beyond. Given the appropriate building blocks being put into 
place to operate within an intelligence architecture (world), the functionality 
you seek would in theory be able to evolve "naturally". Once programmed into a 
functional, matured, intelligent system, this may have potential to serve as a 
future, AGI hub, or docking station. Such is my theory.

According to my analysis, a number of the key components of this intelligence 
architecture have already been programmed by other parties, eg., by CISCO, IBM, 
MS, GOOGLE, OPENCOG, and so many others in the big-wide world. Such progress 
serves to inform me that the overall architecture - formulating in my mind - 
still is on the right track. My passion continues as a hobby. I do it for the 
fun of it. I'm enjoying the journey. My purpose is not to compete with the 
efforts of others, but to indulge my life-long fascination and curiosity with 
the technological challenges presented by AGI, and beyond.

In summary, the one thing needed to solve AGI is, in my opinion, still lacking 
in the public domain.

Rob

From: Alan Grimes 
Sent: Friday, 08 June 2018 2:55 AM
To: Nanograte Knowledge Technologies via AGI
Subject: Re: [agi] The four things needed to solve AGI.

Nanograte Knowledge Technologies via AGI wrote:
> Alan
>
> I suppose yo already possess a computational framework to process the
> research you are proposing as being critical? In other words, I think
> if you do not have the intelligence architecture, none of the
> functionality would be agi effective.


I'm missing a number of things that limits my ability to do this
research myself.

-> a GPU programming framework that actually works on my computer,
preferably OpenCL for portability reasons...

-> a simulation environment... Microsoft Malmo seems pretty close to
what I want (bare minimum actually), I'll have to take a look at it
again to see if I can make it run the way I need it to.

-> Everything is in python these days, so I'll need a high quality
manual for the python language. I don't think this documentation
actually exists. In fairness, I haven't made a concerted effort to
search for it recently. on-line courses in the subject only cover the
whitespace-based nesting but basically ignores all other parts of syntax
such as data definition and expression syntax.

--
Please report bounces from this address to a...@numentics.com

Powers are not rights.

Artificial General Intelligence List / AGI / 
see 

Re: [agi] The four things needed to solve AGI.

[MP via AGI]

Alan, would you be interested in emailing me these models? I’m an experienced 
coder and would love to work on it.

mindpixel at proton mail dot com

Sent from ProtonMail Mobile

On Fri, Jun 8, 2018 at 1:31 AM, Nanograte Knowledge Technologies via AGI 
 wrote:

> Alan
>
> Thank you for your response. I am in no way trying to challenge your personal 
> research, or view. My questions are genuine and AGI related only. I do not 
> agree with the view of some here who defend A.T. Murray's position by saying 
> he at least does something. I agree with your view that it is simply not 
> enough to just make an effort, therefore AGI would result. The true-AGI 
> problem is many-fold more complex than we may even imagine. But all efforts 
> may bring forth new inspiration.
>
> When compared to aspects of AGI being produced in robotic form today, I have 
> no idea why Murray et al are trying to reinvent any one of the wheels that 
> are clearly, already turning. It behooves progress to continually refine the 
> real questions to be addressed, as you have obviously done. Perhaps, what 
> would emerge from those efforts would find practical use?However, for its 
> lack of structured debate between simple and complex systems, I would not be 
> able to relate to it. That's my definitive view on Murray's world.
>
> I do not program anymore, but I have been designing computational theory for 
> system of systems - to purpose - for a long time. AGI, as a holistic 
> architecture, can be nothing less than a galaxy of deliberate system of 
> systems, administered by at least 1, superpositioned system of systems. I 
> have field tested my theoretical designs against real projects and with real 
> people. The results have been accurate and reliable. The mature models are 
> ready for programming.
>
> This is the intelligence (AGI) architecture I am referring to. In my view, it 
> represents as a fully-recursive, adaptive, intelligence architecture (not 
> intelligence bootstrapped), which would go a long way towards enabling AGI 
> functionality, and beyond. Given the appropriate building blocks being put 
> into place to operate within an intelligence architecture (world), the 
> functionality you seek would in theory be able to evolve "naturally". Once 
> programmed into a functional, matured, intelligent system, this may have 
> potential to serve as a future, AGI hub, or docking station. Such is my 
> theory.
>
> According to my analysis, a number of the key components of this intelligence 
> architecture have already been programmed by other parties, eg., by CISCO, 
> IBM, MS, GOOGLE, OPENCOG, and so many others in the big-wide world. Such 
> progress serves to inform me that the overall architecture - formulating in 
> my mind - still is on the right track. My passion continues as a hobby. I do 
> it for the fun of it. I'm enjoying the journey. My purpose is not to compete 
> with the efforts of others, but to indulge my life-long fascination and 
> curiosity with the technological challenges presented by AGI, and beyond.
>
> In summary, the one thing needed to solve AGI is, in my opinion, still 
> lacking in the public domain.
>
> Rob
> ---
>
> From: Alan Grimes 
> Sent: Friday, 08 June 2018 2:55 AM
> To: Nanograte Knowledge Technologies via AGI
> Subject: Re: [agi] The four things needed to solve AGI.
>
> Nanograte Knowledge Technologies via AGI wrote:
>> Alan
>>
>> I suppose yo already possess a computational framework to process the
>> research you are proposing as being critical? In other words, I think
>> if you do not have the intelligence architecture, none of the
>> functionality would be agi effective.
> 
> I'm missing a number of things that limits my ability to do this
> research myself.
> 
> -> a GPU programming framework that actually works on my computer,
> preferably OpenCL for portability reasons...
> 
> -> a simulation environment... Microsoft Malmo seems pretty close to
> what I want (bare minimum actually), I'll have to take a look at it
> again to see if I can make it run the way I need it to.
> 
> -> Everything is in python these days, so I'll need a high quality
> manual for the python language. I don't think this documentation
> actually exists. In fairness, I haven't made a concerted effort to
> search for it recently. on-line courses in the subject only cover the
> whitespace-based nesting but basically ignores all other parts of syntax
> such as data definition and expression syntax.
> 
> --
> Please report bounces from this address to a...@numentics.com
> 
> Powers are not rights.
> [Artificial General Intelligence List](https://agi.topicbox.com/latest) / AGI 
> / see [discussions](https://agi.topicbox.com/groups/agi) + 
> [participants](https://agi.topicbox.com/groups/agi/members) + [delivery 
> options](https://agi.topicbox.com/groups) 
> [Permalink](https://agi.topicbox.com/groups/agi/T507c404b4595c71c-M5774ca24e5367098d2c5ee97)