future of maxentropy module (was: sparse rmatvec and maxentropy)

classic Classic list List threaded Threaded
6 messages Options
Reply | Threaded
Open this post in threaded view
|

future of maxentropy module (was: sparse rmatvec and maxentropy)

Ralf Gommers-2
(excuse the cross-post, but this may be of interest to scipy-user and the scikits.learn crowd)


On Sat, Jan 22, 2011 at 10:44 PM, <[hidden email]> wrote:
On Sat, Jan 22, 2011 at 8:50 AM, Ralf Gommers
<[hidden email]> wrote:
> On Thu, Jan 20, 2011 at 10:13 PM, Skipper Seabold <[hidden email]>
> wrote:
>>
>> I picked up the montecarlo code when I was playing around with these.
>>
>> http://bazaar.launchpad.net/~jsseabold/statsmodels/statsmodels-skipper-maxent/files/head:/scikits/statsmodels/sandbox/maxentropy/
>>
>> I'm curious if the maxentropy stuff as it is in scipy wouldn't find
>> more use and maintenance in scikits.learn.  The implementation is
>> somewhat use specific (natural language processing), though this is
>> not by any means set in stone.
>>
> Probably, but wouldn't it need a lot of work before it could be moved? It
> has a grand total of one test, mostly non-working examples, and is obviously
> hardly used at all (see r6919 and r6920 for more examples of broken code).
>
> Perhaps it's worth asking the scikits.learn guys, and otherwise consider
> deprecating it if they're not interested?

I haven't seen or heard anyone using it besides Skipper. There are
also still some features that where designed for pysparse and never
fully updated to scipy.sparse.
http://projects.scipy.org/scipy/ticket/856

I also thought deprecating and removing maxentropy will be the best
idea, if nobody volunteers to give it a workout.

So I guess we just have to ask this out loud: is anyone using the scipy.maxentropy module or interested in doing so? If you are, would you be interested in putting some work into it, like making the examples work and adding some tests?

The current status is that 3 out of 4 examples are broken, the module has only a single test, and from broken code that went unnoticed for a long time it is clear that there are very few users.

If no one steps up, I propose to deprecate the module for the 0.10 release. If there are any users out there that missed this email and step up then, we can always un-deprecate again.

To the scikits.learn developers: would this code fit better and see more use in scikits.learn than in scipy? Would you be interested to pick it up?

Ralf

_______________________________________________
SciPy-User mailing list
[hidden email]
http://mail.scipy.org/mailman/listinfo/scipy-user
Reply | Threaded
Open this post in threaded view
|

Re: [Scikit-learn-general] future of maxentropy module (was: sparse rmatvec and maxentropy)

Olivier Grisel-3
2011/1/24 Ralf Gommers <[hidden email]>:
>
> To the scikits.learn developers: would this code fit better and see more use
> in scikits.learn than in scipy? Would you be interested to pick it up?

There is already a maxent model in scikit learn which is a wrapper for
LibLinear :

 scikits.learn.linear_model.LogisticRegression

AFAIK, LibLinear is pretty much state of the art so I don't think the
scikits.learn project is interested reusing this code.

Best,

--
Olivier
http://twitter.com/ogrisel - http://github.com/ogrisel
_______________________________________________
SciPy-User mailing list
[hidden email]
http://mail.scipy.org/mailman/listinfo/scipy-user
Reply | Threaded
Open this post in threaded view
|

Re: future of maxentropy module (was: sparse rmatvec and maxentropy)

jseabold
In reply to this post by Ralf Gommers-2
On Mon, Jan 24, 2011 at 9:35 AM, Ralf Gommers
<[hidden email]> wrote:

> (excuse the cross-post, but this may be of interest to scipy-user and the
> scikits.learn crowd)
>
>
> On Sat, Jan 22, 2011 at 10:44 PM, <[hidden email]> wrote:
>>
>> On Sat, Jan 22, 2011 at 8:50 AM, Ralf Gommers
>> <[hidden email]> wrote:
>> > On Thu, Jan 20, 2011 at 10:13 PM, Skipper Seabold <[hidden email]>
>> > wrote:
>> >>
>> >> I picked up the montecarlo code when I was playing around with these.
>> >>
>> >>
>> >> http://bazaar.launchpad.net/~jsseabold/statsmodels/statsmodels-skipper-maxent/files/head:/scikits/statsmodels/sandbox/maxentropy/
>> >>
>> >> I'm curious if the maxentropy stuff as it is in scipy wouldn't find
>> >> more use and maintenance in scikits.learn.  The implementation is
>> >> somewhat use specific (natural language processing), though this is
>> >> not by any means set in stone.
>> >>
>> > Probably, but wouldn't it need a lot of work before it could be moved?
>> > It
>> > has a grand total of one test, mostly non-working examples, and is
>> > obviously
>> > hardly used at all (see r6919 and r6920 for more examples of broken
>> > code).
>> >
>> > Perhaps it's worth asking the scikits.learn guys, and otherwise consider
>> > deprecating it if they're not interested?
>>
>> I haven't seen or heard anyone using it besides Skipper. There are
>> also still some features that where designed for pysparse and never
>> fully updated to scipy.sparse.
>> http://projects.scipy.org/scipy/ticket/856
>>
>> I also thought deprecating and removing maxentropy will be the best
>> idea, if nobody volunteers to give it a workout.
>
> So I guess we just have to ask this out loud: is anyone using the
> scipy.maxentropy module or interested in doing so? If you are, would you be
> interested in putting some work into it, like making the examples work and
> adding some tests?
>
> The current status is that 3 out of 4 examples are broken, the module has
> only a single test, and from broken code that went unnoticed for a long time
> it is clear that there are very few users.
>

I just checked again, and I do have the examples working in
statsmodels with scipy before rmatvec was removed, so it's not so
dire.  It just depends on the montecarlo code, so we would have to
include this in an install if we want the examples to run.  I can make
a branch that includes this code if there's interest to keep it and
have the examples work.

> If no one steps up, I propose to deprecate the module for the 0.10 release.
> If there are any users out there that missed this email and step up then, we
> can always un-deprecate again.
>

I do use things from the code, ie., the scipy.maxentropy.logsumexp, so
I wouldn't want to lose that at the very least.

Skipper
_______________________________________________
SciPy-User mailing list
[hidden email]
http://mail.scipy.org/mailman/listinfo/scipy-user
Reply | Threaded
Open this post in threaded view
|

Re: future of maxentropy module (was: sparse rmatvec and maxentropy)

Ralf Gommers-2


On Mon, Jan 24, 2011 at 11:03 PM, Skipper Seabold <[hidden email]> wrote:
On Mon, Jan 24, 2011 at 9:35 AM, Ralf Gommers
<[hidden email]> wrote:
> (excuse the cross-post, but this may be of interest to scipy-user and the
> scikits.learn crowd)
>
>
> On Sat, Jan 22, 2011 at 10:44 PM, <[hidden email]> wrote:
>>
>> On Sat, Jan 22, 2011 at 8:50 AM, Ralf Gommers
>> <[hidden email]> wrote:
>> > On Thu, Jan 20, 2011 at 10:13 PM, Skipper Seabold <[hidden email]>
>> > wrote:
>> >>
>> >> I picked up the montecarlo code when I was playing around with these.
>> >>
>> >>
>> >> http://bazaar.launchpad.net/~jsseabold/statsmodels/statsmodels-skipper-maxent/files/head:/scikits/statsmodels/sandbox/maxentropy/
>> >>
>> >> I'm curious if the maxentropy stuff as it is in scipy wouldn't find
>> >> more use and maintenance in scikits.learn.  The implementation is
>> >> somewhat use specific (natural language processing), though this is
>> >> not by any means set in stone.
>> >>
>> > Probably, but wouldn't it need a lot of work before it could be moved?
>> > It
>> > has a grand total of one test, mostly non-working examples, and is
>> > obviously
>> > hardly used at all (see r6919 and r6920 for more examples of broken
>> > code).
>> >
>> > Perhaps it's worth asking the scikits.learn guys, and otherwise consider
>> > deprecating it if they're not interested?
>>
>> I haven't seen or heard anyone using it besides Skipper. There are
>> also still some features that where designed for pysparse and never
>> fully updated to scipy.sparse.
>> http://projects.scipy.org/scipy/ticket/856
>>
>> I also thought deprecating and removing maxentropy will be the best
>> idea, if nobody volunteers to give it a workout.
>
> So I guess we just have to ask this out loud: is anyone using the
> scipy.maxentropy module or interested in doing so? If you are, would you be
> interested in putting some work into it, like making the examples work and
> adding some tests?
>
> The current status is that 3 out of 4 examples are broken, the module has
> only a single test, and from broken code that went unnoticed for a long time
> it is clear that there are very few users.
>

I just checked again, and I do have the examples working in
statsmodels with scipy before rmatvec was removed, so it's not so
dire.  It just depends on the montecarlo code, so we would have to
include this in an install if we want the examples to run.  I can make
a branch that includes this code if there's interest to keep it and
have the examples work.

The montecarlo code was removed for a reason I assume, so that would be even more work to include again....

On the scikits.learn list someone said the maxentropy examples are nice, so perhaps they could be made to work with (translated to) the logistic regression code in scikits.learn.

> If no one steps up, I propose to deprecate the module for the 0.10 release.
> If there are any users out there that missed this email and step up then, we
> can always un-deprecate again.
>

I do use things from the code, ie., the scipy.maxentropy.logsumexp, so
I wouldn't want to lose that at the very least.

That's a 3-line long utility function, I'm sure a place could be found for it. Anyway, I'm not proposing to throw the code out tomorrow - after 0.10 is out for a while we could go through it and move anything useful.

Cheers,
Ralf


_______________________________________________
SciPy-User mailing list
[hidden email]
http://mail.scipy.org/mailman/listinfo/scipy-user
Reply | Threaded
Open this post in threaded view
|

Re: [SciPy-Dev] future of maxentropy module (was: sparse rmatvec and maxentropy)

Gael Varoquaux
On Tue, Jan 25, 2011 at 06:15:22PM +0800, Ralf Gommers wrote:
>    On the scikits.learn list someone said the maxentropy examples are nice,
>    so perhaps they could be made to work with (translated to) the logistic
>    regression code in scikits.learn.

OK, I'll see what we can do.

I had a quick look at the examples, and they seemed so synthetic that I
couldn't get the point. But then again, I am not a Natural Language
Processing guy, so I'll see if I can get an NLP guy translate (and
explain) the examples to the scikit.

Ga�l

_______________________________________________
SciPy-User mailing list
[hidden email]
http://mail.scipy.org/mailman/listinfo/scipy-user
Reply | Threaded
Open this post in threaded view
|

Re: [SciPy-Dev] future of maxentropy module (was: sparse rmatvec and maxentropy)

Ralf Gommers-2


On Wed, Jan 26, 2011 at 5:08 AM, Gael Varoquaux <[hidden email]> wrote:
On Tue, Jan 25, 2011 at 06:15:22PM +0800, Ralf Gommers wrote:
>    On the scikits.learn list someone said the maxentropy examples are nice,
>    so perhaps they could be made to work with (translated to) the logistic
>    regression code in scikits.learn.

OK, I'll see what we can do.

I had a quick look at the examples, and they seemed so synthetic that I
couldn't get the point. But then again, I am not a Natural Language
Processing guy, so I'll see if I can get an NLP guy translate (and
explain) the examples to the scikit.

That would be great, thanks. If it's not possible or useful, an answer from an expert explaining why would also be helpful.

Ralf



_______________________________________________
SciPy-User mailing list
[hidden email]
http://mail.scipy.org/mailman/listinfo/scipy-user