Preschool: sketch of an evaluation framework for early stage AGI systems
aimed at human-level, roughly humanlike AGI
To: agi@v2.listbox.com
Date: Sunday, December 28, 2008, 11:38 PM
Robert,
Thanks for your detailed, helpful replies. I like your approach of operating in
multiple domains
The paper as a link instead of attachment:
http://mindsoftbioware.com/yahoo_site_admin/assets/docs/Swaine_R_Story_Understander_Model.36375123.pdf
The paper gives a quick view of the Human-centric representation and behavioral
systems approach for problem-solving, reasoning as giving meaning
Robert,
What kind of problems have you designed this to solve? Can you give some
examples?
Robert:
A brief paper on an AGI system for human-level ...had only 2 pages to
fit in.
If you are working on a system, you probably hope it will one day help
design a better world,
]
AGI Preschool: sketch of an evaluation framework for early stage AGI systems
aimed at human-level, roughly humanlike AGI
To: agi@v2.listbox.com
Date: Sunday, December 28, 2008, 4:49 PM
Robert,
What kind of problems have you designed this to solve? Can you give some
examples?
Robert:
A brief
Robert:
Example:
Here's a pattern example you may not have seen before, but by 3C you discover
the pattern and how to make an example:
As spoken aloud:
five and nine[is] fine
two and six [is] twix
five and seven [is] fiven
Robert,
So, if I understand, you're designing a system
approach-paper (was Re: Indexing and Re: [agi]
AGI Preschool: sketch of an evaluation framework for early stage AGI systems
aimed at human-level, roughly humanlike AGI
To: agi@v2.listbox.com
Date: Sunday, December 28, 2008, 8:38 PM
Robert:
Example:
Here's a pattern example you may not have
approach-paper (was Re: Indexing and
Re: [agi] AGI Preschool: sketch of an evaluation framework for early stage AGI
systems aimed at human-level, roughly humanlike AGI
To: agi@v2.listbox.com
Date: Sunday, December 28, 2008, 8:38 PM
Robert:
Example
On Sat, Dec 20, 2008 at 8:01 AM, Derek Zahn derekz...@msn.com wrote:
Ben:
Right. My intuition is that we don't need to simulate the dynamics
of fluids, powders and the like in our virtual world to make it adequate
for teaching AGIs humanlike, human-level AGI. But this could be
wrong.
It's an interesting idea, but I suspect it too will rapidly break down.
Which activities can be known about in a rich, better-than-blind-Cyc way
*without* a knowledge of objects and object manipulation? How can an agent
know about reading a book,for example, if it can't pick up and
I agree, but the good news is that game dev advances fast.
So, my plan with the AGI Preschool would be to build it in an open platform
such as OpenSim, and then swap in better and better physics engines as they
become available.
Some current robot simulators use ODE and this seems to be good
Some might say that if they get conservation of mass
and newton's law then they skipped all the useless stuff!
OK, but those some probably don't include any preschool
teachers or educational theorists. That hypothesis is completely at odds
with my own intuition
from having raised 3
2008/12/20 Ben Goertzel b...@goertzel.org:
Well, it's completely obvious to me, based on my knowledge of virtual worlds
and robotics, that building a high quality virtual world is orders of
magnitude easier than making a workable humanoid robot.
I guess that depends on what you mean by high
Ben: Right. My intuition is that we don't need to simulate the dynamics of
fluids, powders and the like in our virtual world to make it adequate for
teaching AGIs humanlike, human-level AGI. But this could be wrong.I suppose
it depends on what kids actually learn when making cakes, skipping
On Sat, Dec 20, 2008 at 10:44 AM, Philip Hunt cabala...@googlemail.comwrote:
2008/12/20 Ben Goertzel b...@goertzel.org:
Well, it's completely obvious to me, based on my knowledge of virtual
worlds
and robotics, that building a high quality virtual world is orders of
magnitude easier
Oh, and because I am interested in the potential of high-fidelity physical
simulation as a basis for AI research, I did spend some time recently looking
into options. Unfortunately the results, from my perspective, were
disappointing.
The common open-source physics libraries like ODE,
Bob: Even with crude or no real
simulation ability in an environment such as Second Life, using some
simple symbology to stand for puck up screwdriver you can still try
to tackle problems such as autobiographical memory - how does the
agent create a coherent story out of a series of activities,
2008/12/20 Ben Goertzel b...@goertzel.org:
It doesn't have to be humanoid ... but apart from rolling instead of
walking,
I don't see any really significant simplifications obtainable from making it
non-humanoid.
I can think of several. For example, you could give it lidar to
measure
2008/12/20 Derek Zahn derekz...@msn.com:
Ben:
Right. My intuition is that we don't need to simulate the dynamics
of fluids, powders and the like in our virtual world to make it adequate
for teaching AGIs humanlike, human-level AGI. But this could be
wrong.
I suppose it depends on what
Well, there is massively more $$ going into robotics dev than into AGI dev,
and no one seems remotely near to solving the hard problems
Which is not to say it's a bad area of research, just that it's a whole
other huge confusing RD can of worms
So I still say, the choices are
-- virtual
Consider an object, such as a sock or a book or a cat. These objects
can all be recognised by young children, even though the visual input
coming from trhem chasnges from what angle they're viewed at. More
fundamentally, all these objects can change shape, yet humans can
still effortlessly
2008/12/20 Ben Goertzel b...@goertzel.org:
However, I have long been perplexed at the obsession with so many AI folks
with vision processing.
I wouldn't say I'm obsessed with it. On its own vision processing
does nothing, the same as all other input processing -- its only when
a brain/AI used
A paper by Stephan Bugaj and I that will appear in the AGI-09 proceedings
and get presented at the conference.
http://www.opencog.org/wiki/Image:Preschool.pdf
I'll also be giving a couple technical papers together w/ other colleagues,
but this one focuses on how to evaluate AGIs and so may be of
2008/12/19 Ben Goertzel b...@goertzel.org:
What I'd like to see is a really nicely implemented virtual world
preschool for AIs ... though of course building such a thing will be a lot
of work for someone...
Why a virtual world preschool and not a real one?
A virtual world, if not programmed
Well, there is a major question whether one can meaningfully address AGI via
virtual-robotics rather than physical-robotics
No one can make a convincing proof either way right now
But, it's clear that if one wants to go the physical-robotics direction, now
is not the time to be working on
Hi Ben.
OTOH, if one wants to go the virtual-robotics direction (as is my intuition),
then it is possible to bypass many of the lower-level perception/actuation
issues and focus on preschool-level learning, reasoning and conceptual
creation.
And yet, in your paper (which I enjoyed),
It's a hard problem, and the answer is to cheat as much as possible, but
not any more so.
We'll just have to feel this out via experiment...
My intuition is that current virtual worlds and game worlds are too crude,
but current robot simulators are not.
I.e., I doubt one needs serious fluid
2008/12/20 Ben Goertzel b...@goertzel.org:
I.e., I doubt one needs serious fluid dynamics in one's simulation ... I
doubt one needs bodies with detailed internal musculature ... but I think
one does need basic Newtonian physics and the ability to use tools, break
things in half (but not
On Fri, Dec 19, 2008 at 8:42 PM, Philip Hunt cabala...@googlemail.comwrote:
2008/12/20 Ben Goertzel b...@goertzel.org:
I.e., I doubt one needs serious fluid dynamics in one's simulation ... I
doubt one needs bodies with detailed internal musculature ... but I think
one does need basic
2008/12/20 Derek Zahn derekz...@msn.com:
And yet, in your paper (which I enjoyed), you emphasize the importance of
not providing
a simplistic environment (with the screwdriver example). Without facing the
low-level
sensory world (either through robotics or through very advanced simulations
2008/12/20 Ben Goertzel b...@goertzel.org:
3. to provide a toy domain for the AI to think about and become
proficient in.
Not just to become proficient in the domain, but become proficient
in general humanlike cognitive processes.
The point of a preschool is that it's designed to present
Right. My intuition is that we don't need to simulate the dynamics
of fluids, powders and the like in our virtual world to make it adequate
for teaching AGIs humanlike, human-level AGI. But this could be
wrong.
It also could be interesting to program an artificial chemistry that
emulated
Well, that's a really easy example, right? For making tea, the answer
would probably be yes.
Baking a cake is a harder example. An AGI trained in a virtual world could
certainly follow a recipe to make a passable cake. But it would never learn
to be a **really good** baker in the virtual
2008/12/20 Ben Goertzel b...@goertzel.org:
Baking a cake is a harder example. An AGI trained in a virtual world could
certainly follow a recipe to make a passable cake. But it would never learn
to be a **really good** baker in the virtual world, unless the virtual world
were fabulously
Ahhh... ***that's*** why everyone always hates my cakes!!! I never realized
you were supposed to **taste** the stuff ... I thought it was just supposed
to look funky after you throw it in somebody's face ;-)
On Fri, Dec 19, 2008 at 9:31 PM, Philip Hunt cabala...@googlemail.comwrote:
2008/12/20
Although, I note, I know a really good baker who makes great cakes in spite
of the fact that she does not eat sugar and hence does not ever taste most
of the stuff she makes...
But she *used to* eat sugar, so to an extent she can go on memory
Sorta like how Beethoven kept composing after he went
On Dec 19, 2008, at 6:43 PM, Ben Goertzel wrote:
Although, I note, I know a really good baker who makes great cakes
in spite of the fact that she does not eat sugar and hence does not
ever taste most of the stuff she makes...
But she *used to* eat sugar, so to an extent she can go on
36 matches
Mail list logo
