Solving Word Problems with Sage

[Timothy Clemans]

Hi,

I created a rudimentary word problem solver, see
http://sagenb.org/home/pub/1096/ What improvements are needed in order
for it be seriously considered for inclusion in Sage?

Thanks,
Timothy

[William Stein]

"Question: What is the cube root of 27? Answer: The cube root of 27
is 9." [[oops!]]

Are you trying to create a competitor to Worlfram Alpha? I hope so :-)

-- William

>
> Thanks,
> Timothy
>
> --
>
> You received this message because you are subscribed to the Google Groups "sage-edu" group.
> To post to this group, send email to sage...@googlegroups.com.
> To unsubscribe from this group, send email to sage-edu+u...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/sage-edu?hl=.
>
>
>



--
William Stein
Associate Professor of Mathematics
University of Washington
http://wstein.org

[Timothy Clemans]

On Fri, Nov 20, 2009 at 12:14 AM, William Stein <wst...@gmail.com> wrote:
> On Fri, Nov 20, 2009 at 12:07 AM, Timothy Clemans
> <timothy...@gmail.com> wrote:
>> Hi,
>>
>> I created a rudimentary word problem solver, see
>> http://sagenb.org/home/pub/1096/ What improvements are needed in order
>> for it be seriously considered for inclusion in Sage?
>
> "Question: What is the cube root of 27? Answer: The cube root of 27
> is 9."  [[oops!]]
>
> Are you trying to create a competitor to Worlfram Alpha?   I hope so :-)

Yes I am trying to create a competitor to Wolfram Alpha.

[jason...@creativetrax.com]

Timothy Clemans wrote:
> On Fri, Nov 20, 2009 at 12:14 AM, William Stein <wst...@gmail.com> wrote:
>
>> On Fri, Nov 20, 2009 at 12:07 AM, Timothy Clemans
>> <timothy...@gmail.com> wrote:
>>
>>> Hi,
>>>
>>> I created a rudimentary word problem solver, see
>>> http://sagenb.org/home/pub/1096/ What improvements are needed in order
>>> for it be seriously considered for inclusion in Sage?
>>>
>> "Question: What is the cube root of 27? Answer: The cube root of 27
>> is 9." [[oops!]]
>>
>> Are you trying to create a competitor to Worlfram Alpha? I hope so :-)
>>
>
> Yes I am trying to create a competitor to Wolfram Alpha.
>
>

Here's maybe a relevant link: a natural language processing package for
python:

http://www.nltk.org/

Jason

[jason...@creativetrax.com]

(It's amazing how much quality stuff we get for free because we are
open-source and use python as a core language!)

Jason

[William Stein]

On Fri, Nov 20, 2009 at 1:01 AM, <jason...@creativetrax.com> wrote:
> jason...@creativetrax.com wrote:
>> Here's maybe a relevant link: a natural language processing package for
>> python:
>>
>> http://www.nltk.org/
>>
>
>
> (It's amazing how much quality stuff we get for free because we are
> open-source and use python as a core language!)
>
> Jason

We definitely get a *lot* more than Wolfram gets.

-- William

[Timothy Clemans]

What should the first version of Sage Answers be able to do?

[David Joyner]

On Fri, Nov 20, 2009 at 4:15 AM, Timothy Clemans
<timothy...@gmail.com> wrote:
> What should the first version of Sage Answers be able to do?
>

Just to begin, you could try to set it up so that it can solve basic problems
that a freshman calculus student might ask. For example, the
plot/integral/derivative
of a simple function, or the root of a simple equation.

[Jason Grout]

David Joyner wrote:
> On Fri, Nov 20, 2009 at 4:15 AM, Timothy Clemans
> <timothy...@gmail.com> wrote:
>> What should the first version of Sage Answers be able to do?
>>
>
> Just to begin, you could try to set it up so that it can solve basic problems
> that a freshman calculus student might ask. For example, the
> plot/integral/derivative
> of a simple function, or the root of a simple equation.
>
>

I think most of my students use Alpha for stuff like:

integrate ___ from __ to __

plot ___

differentiate ___ with respect to ___

etc.


Or really, they just plop the function in and wolfram automatically
plots it.

Jason

[William Stein]

Hi,

According to those articles posted on the Wolfram blog, Wolfram Alpha
is "over 7 million lines of code" and maybe a decade of work by a lot
of people. I wonder if one saw the actual queries that are made, if
*most* of them (say the top 95%), then it would be relatively easy to
implement something that could answer those top 95% of questions?
In the Wolfram blog he particularly talks about how they had to go
public with Wolfram Alpha since they desperately wanted to see what
people would query for, and without that information they really
couldn't go further with the project.

So... Timothy, is there any way you can find out anything about the
top most popular Wolfram Alpha queries? Also, it would be worth
looking at the top queries at http://mathoverflow.net/ and other
similar sites (though maybe those are likely to be a lot more
complicated?)

William