Randomly generated progressive matrices game?
608 views
Skip to first unread message
derpaderp
Apr 27, 2012, 3:40:37 AM4/27/12
to Dual N-Back, Brain Training & Intelligence
I've been searching for something like this for a couple of days with
no luck. It seems strange to me that this hasn't been done yet. The
closest thing I've found is the object reasoning game at the cambridge
brain sciences site, which is complete garbage.
no luck. It seems strange to me that this hasn't been done yet. The
closest thing I've found is the object reasoning game at the cambridge
brain sciences site, which is complete garbage.
☉
Apr 27, 2012, 9:41:47 PM4/27/12
to Dual N-Back, Brain Training & Intelligence
It would appear you might be asking for something like the Sandia
Generated Matrix Tool: http://cognitivescience.sandia.gov/Software/matrixtool.html
. I personally find it to be quite ugly and rather uninteresting in
the kinds of problems it generates.
There is also the more game-like task called the Odd One Out (besides
Object Reasoning) available at the Cambridge Brain Sciences website,
but it also has the severe (damming) defect of being easily hacked
while not being completely of the "progressive matrix" variety.
(Actually, this kind of problem subsumes progressive matrices, because
you're given more information in the former than the latter.)
Other than the SGMT, I'm not sure what else might be out there.
argumzio
Generated Matrix Tool: http://cognitivescience.sandia.gov/Software/matrixtool.html
. I personally find it to be quite ugly and rather uninteresting in
the kinds of problems it generates.
There is also the more game-like task called the Odd One Out (besides
Object Reasoning) available at the Cambridge Brain Sciences website,
but it also has the severe (damming) defect of being easily hacked
while not being completely of the "progressive matrix" variety.
(Actually, this kind of problem subsumes progressive matrices, because
you're given more information in the former than the latter.)
Other than the SGMT, I'm not sure what else might be out there.
argumzio
polar
Apr 28, 2012, 7:17:32 AM4/28/12
to brain-t...@googlegroups.com
Looks nice, but I think this software should not be publicly available - it already contains some of the RAPM problem types, and as a open source it can be enhanced with additional problem types. I am well aware that lot of things can be considered a iq-test-taking training method, but it can hardly get closer than this. If anyone wants to train complex, wm taxing reasoning, I think very close to most advanced RAPM problems is the "einstein's riddle" ( http://www.kongregate.com/games/sheldonx/einsteins-riddle ). I even had a downloadable game which generated this type of tasks at will, but now I cant find it - anybody have it by chance?
Dne sobota, 28. dubna 2012 3:41:47 UTC+2 ☉ napsal(a):
Dne sobota, 28. dubna 2012 3:41:47 UTC+2 ☉ napsal(a):
☉
Apr 28, 2012, 1:13:51 PM4/28/12
to Dual N-Back, Brain Training & Intelligence
Yes, I agree about the SGMT, but the RPM-type problems as they
generally exist now are known pretty much anywhere online.
Anyway, I don't really think the "Einstein puzzle" has much to do with
matrix reasoning, as the latter is more about pattern recognition, and
has very little cultural load (i.e., requirement for prior knowledge).
Of course, it would be easy to make such a puzzle with abstract
symbols, but at that point it still doesn't manifest the rich variety
of needed reasoning and pattern recognition. That is, the Einstein
puzzle problem is really a mechanical process, and I don't think it
will do much for matrix reasoning problems like Raven's. It is a good
evaluator of WMC, but only up to a point, I suspect.
Here's one source I found: http://www.logic-puzzles.org/index.php . I
also remember a downloadable version of such a program (which also
used abstract symbols), but I have forgotten what it's called, as it
wasn't that interesting to me.
argumzio
On Apr 28, 6:17 am, polar <pol...@gmail.com> wrote:
> Looks nice, but I think this software should not be publicly available - it
> already contains some of the RAPM problem types, and as a open source it
> can be enhanced with additional problem types. I am well aware that lot of
> things can be considered a iq-test-taking training method, but it can
> hardly get closer than this. If anyone wants to train complex, wm taxing
> reasoning, I think very close to most advanced RAPM problems is the
> "einstein's riddle" (http://www.kongregate.com/games/sheldonx/einsteins-riddle). I even had a
generally exist now are known pretty much anywhere online.
Anyway, I don't really think the "Einstein puzzle" has much to do with
matrix reasoning, as the latter is more about pattern recognition, and
has very little cultural load (i.e., requirement for prior knowledge).
Of course, it would be easy to make such a puzzle with abstract
symbols, but at that point it still doesn't manifest the rich variety
of needed reasoning and pattern recognition. That is, the Einstein
puzzle problem is really a mechanical process, and I don't think it
will do much for matrix reasoning problems like Raven's. It is a good
evaluator of WMC, but only up to a point, I suspect.
Here's one source I found: http://www.logic-puzzles.org/index.php . I
also remember a downloadable version of such a program (which also
used abstract symbols), but I have forgotten what it's called, as it
wasn't that interesting to me.
argumzio
On Apr 28, 6:17 am, polar <pol...@gmail.com> wrote:
> Looks nice, but I think this software should not be publicly available - it
> already contains some of the RAPM problem types, and as a open source it
> can be enhanced with additional problem types. I am well aware that lot of
> things can be considered a iq-test-taking training method, but it can
> hardly get closer than this. If anyone wants to train complex, wm taxing
> reasoning, I think very close to most advanced RAPM problems is the
polar
Apr 29, 2012, 6:27:43 AM4/29/12
to brain-t...@googlegroups.com
No big deal, but our culture is not culture load for us, and rapm tasks are not only pattern recognition, but pattern application too (you have to check all the time if your ideas are working). In addition, I would not reduce pattern application (or the riddle) to a mechanical process - e.g. low level sudoku is mechanical, high level sudoku needs you to take "further irreducible" leaps of reasoning in your mind.
Dne sobota, 28. dubna 2012 19:13:51 UTC+2 ☉ napsal(a):
Reply all
Reply to author
Forward
0 new messages