tuxmath-devel

[Tuxmath-devel] Personalized, adaptive game-play to improve learning in Tuxmath

[Siddharth K. <sid...@gm...>]

Hi everyone,

If the devs. of Tuxmath agree, I plan to submit a SoC proposal on the idea
of providing "an adaptive game-play to improve learning" in Tuxmath under
PSU <http://summer.cs.pdx.edu/ideas>. The reason for choosing PSU is that
they focus on academic-oriented projects and consider how valuable the the
project could be to the larger open source community.

*Short description*: Tux, of Math command is an open-source educational game
for learning Mathematics. Since the game is quite popular and used by many
schools, my idea focuses on improving the learning of a player/student by
providing a personalized, adaptive game-play. This is done by modelling the
student's knowledge state using a Bayesian network. This also helps a
teacher/supervisor to be provided with a detailed assessment about a
particular student's strengths and weaknesses in individual concepts (read
lessons).

If this idea adds value to the tux community, I would like to work on it
irrespective of whether I make it into SoC, though time would be a
constraint. Do let me know your opinions.

Thanks,
Siddharth

Re: [Tuxmath-devel] Personalized, adaptive game-play to improve learning in Tuxmath

[Tim H. <ho...@wu...>]

Hi Siddarth,

On Thursday, April 07, 2011 10:21:36 am Siddharth Kothari wrote:
> If the devs. of Tuxmath agree, I plan to submit a SoC proposal on the idea
> of providing "an adaptive game-play to improve learning" in Tuxmath under
> PSU <http://summer.cs.pdx.edu/ideas>. The reason for choosing PSU is that
> they focus on academic-oriented projects and consider how valuable the the
> project could be to the larger open source community.
> 
> *Short description*: Tux, of Math command is an open-source educational
> game for learning Mathematics. Since the game is quite popular and used by
> many schools, my idea focuses on improving the learning of a
> player/student by providing a personalized, adaptive game-play. This is
> done by modelling the student's knowledge state using a Bayesian network.
> This also helps a teacher/supervisor to be provided with a detailed
> assessment about a particular student's strengths and weaknesses in
> individual concepts (read lessons).

I think this is an interesting idea. Tuxmath currently has a form of 
"feedback," adjusting the comet speed depending on how the student has 
performed recently. But it is very primitive (I can say that because I wrote 
it myself :-) ). There is plenty of room for improvement, and the idea of 
using Bayesian methods has potential.

Currently, I think the general principle is that:
1. Adaptive mechanisms are desirable in the "training academy" lessons;
2. The arcade games are supposed to be an absolute scale, and so there should 
not be any mechanisms that adjust task difficulty. In principle, you could get 
around this by adjusting how tuxmath does scoring, but I would be very 
cautious about doing this until you have considerable confidence that you can 
measure performance accurately. So you would want to disable any fancy 
mechanisms in the arcade games.

Another thing you should know about is the tux4kids-admin project, which is 
designed to build infrastructure for teachers to use tux4kids in more of a 
"lesson" environment. I implemented a first stab at this, but it never got to 
the point of being very user-friendly. There has been a more recent effort at 
it, but I am not sure of its current status. But you may want to take a peek 
at that work to see how your thoughts might relate.

In any event, I like this general direction of investigation, but I cannot 
promise to be able to supply mentoring time---I haven't had time for tux4kids 
for far longer than I'd like, and so I won't make any promises.

Best,
--Tim

Re: [Tuxmath-devel] Personalized, adaptive game-play to improve learning in Tuxmath

[Siddharth K. <sid...@gm...>]

Hi Tim and everyone,

2. The arcade games are supposed to be an absolute scale, and so there
> should
> not be any mechanisms that adjust task difficulty. In principle, you could
> get
> around this by adjusting how tuxmath does scoring, but I would be very
> cautious about doing this until you have considerable confidence that you
> can
> measure performance accurately. So you would want to disable any fancy
> mechanisms in the arcade games.
>
I do not intend to adjust task difficulty. My initial idea is to create
dependencies between the topics, like multiplication requires knowledge of
addition; but on a more *granular *basis by making distinction between
different types of multiplication, based on the digits in the answer, the
digits in the operands, and also the position of the operands.


> Another thing you should know about is the tux4kids-admin project, which is
> designed to build infrastructure for teachers to use tux4kids in more of a
> "lesson" environment. I implemented a first stab at this, but it never got
> to
> the point of being very user-friendly. There has been a more recent effort
> at
> it, but I am not sure of its current status. But you may want to take a
> peek
> at that work to see how your thoughts might relate.
>
I have used tux4kids-admin tool, but there is not enough documentation and I
haven't gone through the code to completely understand what parts are still
missing. Imo, my idea would relate to this module by allowing assessment of
student's performance in a more accurate manner.

Link to my proposal:
http://www.google-melange.com/gsoc/proposal/review/google/gsoc2011/sidi/2001.
Feedbacks welcome.

Thanks,
Siddharth
On Thu, Apr 7, 2011 at 9:14 PM, Tim Holy <ho...@wu...> wrote:

> Hi Siddarth,
>
> On Thursday, April 07, 2011 10:21:36 am Siddharth Kothari wrote:
> > If the devs. of Tuxmath agree, I plan to submit a SoC proposal on the
> idea
> > of providing "an adaptive game-play to improve learning" in Tuxmath under
> > PSU <http://summer.cs.pdx.edu/ideas>. The reason for choosing PSU is
> that
> > they focus on academic-oriented projects and consider how valuable the
> the
> > project could be to the larger open source community.
> >
> > *Short description*: Tux, of Math command is an open-source educational
> > game for learning Mathematics. Since the game is quite popular and used
> by
> > many schools, my idea focuses on improving the learning of a
> > player/student by providing a personalized, adaptive game-play. This is
> > done by modelling the student's knowledge state using a Bayesian network.
> > This also helps a teacher/supervisor to be provided with a detailed
> > assessment about a particular student's strengths and weaknesses in
> > individual concepts (read lessons).
>
> I think this is an interesting idea. Tuxmath currently has a form of
> "feedback," adjusting the comet speed depending on how the student has
> performed recently. But it is very primitive (I can say that because I
> wrote
> it myself :-) ). There is plenty of room for improvement, and the idea of
> using Bayesian methods has potential.
>
> Currently, I think the general principle is that:
> 1. Adaptive mechanisms are desirable in the "training academy" lessons;
> 2. The arcade games are supposed to be an absolute scale, and so there
> should
> not be any mechanisms that adjust task difficulty. In principle, you could
> get
> around this by adjusting how tuxmath does scoring, but I would be very
> cautious about doing this until you have considerable confidence that you
> can
> measure performance accurately. So you would want to disable any fancy
> mechanisms in the arcade games.
>
> Another thing you should know about is the tux4kids-admin project, which is
> designed to build infrastructure for teachers to use tux4kids in more of a
> "lesson" environment. I implemented a first stab at this, but it never got
> to
> the point of being very user-friendly. There has been a more recent effort
> at
> it, but I am not sure of its current status. But you may want to take a
> peek
> at that work to see how your thoughts might relate.
>
> In any event, I like this general direction of investigation, but I cannot
> promise to be able to supply mentoring time---I haven't had time for
> tux4kids
> for far longer than I'd like, and so I won't make any promises.
>
> Best,
> --Tim
>
>
> ------------------------------------------------------------------------------
> Xperia(TM) PLAY
> It's a major breakthrough. An authentic gaming
> smartphone on the nation's most reliable network.
> And it wants your games.
> http://p.sf.net/sfu/verizon-sfdev
> _______________________________________________
> Tuxmath-devel mailing list
> Tux...@li...
> https://lists.sourceforge.net/lists/listinfo/tuxmath-devel
>

Re: [Tuxmath-devel] Personalized, adaptive game-play to improve learning in Tuxmath

[Caroline F. <car...@gm...>]

I thought we didn't make the summer of code this year?

Caroline

Sent from a mobile device.

On 7 Apr 2011, at 16:44, Tim Holy <ho...@wu...> wrote:

> Hi Siddarth,
> 
> On Thursday, April 07, 2011 10:21:36 am Siddharth Kothari wrote:
>> If the devs. of Tuxmath agree, I plan to submit a SoC proposal on the idea
>> of providing "an adaptive game-play to improve learning" in Tuxmath under
>> PSU <http://summer.cs.pdx.edu/ideas>. The reason for choosing PSU is that
>> they focus on academic-oriented projects and consider how valuable the the
>> project could be to the larger open source community.
>> 
>> *Short description*: Tux, of Math command is an open-source educational
>> game for learning Mathematics. Since the game is quite popular and used by
>> many schools, my idea focuses on improving the learning of a
>> player/student by providing a personalized, adaptive game-play. This is
>> done by modelling the student's knowledge state using a Bayesian network.
>> This also helps a teacher/supervisor to be provided with a detailed
>> assessment about a particular student's strengths and weaknesses in
>> individual concepts (read lessons).
> 
> I think this is an interesting idea. Tuxmath currently has a form of 
> "feedback," adjusting the comet speed depending on how the student has 
> performed recently. But it is very primitive (I can say that because I wrote 
> it myself :-) ). There is plenty of room for improvement, and the idea of 
> using Bayesian methods has potential.
> 
> Currently, I think the general principle is that:
> 1. Adaptive mechanisms are desirable in the "training academy" lessons;
> 2. The arcade games are supposed to be an absolute scale, and so there should 
> not be any mechanisms that adjust task difficulty. In principle, you could get 
> around this by adjusting how tuxmath does scoring, but I would be very 
> cautious about doing this until you have considerable confidence that you can 
> measure performance accurately. So you would want to disable any fancy 
> mechanisms in the arcade games.
> 
> Another thing you should know about is the tux4kids-admin project, which is 
> designed to build infrastructure for teachers to use tux4kids in more of a 
> "lesson" environment. I implemented a first stab at this, but it never got to 
> the point of being very user-friendly. There has been a more recent effort at 
> it, but I am not sure of its current status. But you may want to take a peek 
> at that work to see how your thoughts might relate.
> 
> In any event, I like this general direction of investigation, but I cannot 
> promise to be able to supply mentoring time---I haven't had time for tux4kids 
> for far longer than I'd like, and so I won't make any promises.
> 
> Best,
> --Tim
> 
> ------------------------------------------------------------------------------
> Xperia(TM) PLAY
> It's a major breakthrough. An authentic gaming
> smartphone on the nation's most reliable network.
> And it wants your games.
> http://p.sf.net/sfu/verizon-sfdev
> _______________________________________________
> Tuxmath-devel mailing list
> Tux...@li...
> https://lists.sourceforge.net/lists/listinfo/tuxmath-devel

Re: [Tuxmath-devel] Personalized, adaptive game-play to improve learning in Tuxmath

[Brendan L. <bm...@ri...>]

On Thu, Apr 7, 2011 at 11:51 AM, Caroline Ford <car...@gm...
> wrote:

> I thought we didn't make the summer of code this year?
>

That's correct. This year will be the Tux4Kids Summer of Integration and
Cleanup :/

-Brendan