Hi Andy,
Thanks for your reply, and you raise two valid points.
1) I am not familiar with it and therefore have no way of judging the
content. Maybe some other core devs do?
Understandable, but rather than assuring you that Roger and I have selected
high-quality content and intertwined the concepts appropriately (which
would be of little value coming from me), I'll point you to two classes
that have recently used metacademy:
http://www.seas.harvard.edu/courses/cs281/
http://www.genome.duke.edu/labs/engelhardt/courses/sta561.html
2) I am not sure how large the benefit would be. The documentation contains
"see also" and "references" sections. You site basically seems to provide
(very) advanced versions of these. From a user perspective, this seems to
be an additional indirection. So if a user reads about a method, it seems
easier to just give "Elements of statistical learning" as a reference
instead of linking to metacademy, which then gives the reference.
Good point, and I think if you have a specific reference that you'd like
the reader to examine then you should directly cite it rather than
indirectly citing metacademy, e.g. algorithm Y was invented by Z [citation]
<-- should not use metacademy. On the other hand, if you'd like to direct
the reader to "learn more about this concept," then metacademy has a few
advantages over a direct reference: (1) the resources for the concepts have
different levels of expertism, which can appeal to the different
backgrounds and goals of the reader, (2) the concepts have an associated
prerequisite structure, so for instance, a reader could see that
understanding bayesian networks is important to understanding HMMs, (3)
when possible, our references directly link to the desired location within
larger documents: so from your example, adaboost in the skl
documentation<http://scikit-learn.org/stable/modules/ensemble.html#adaboost>directs
the reader to ESl while adaboost
in metacademy <http://metacademy.org/graphs/concepts/ada_boost> directly
links to section 10.1 "Boosting Methods" in ESL (it also has direct links
to the coursera videos on the topic, which provide a gentler presentation).
Thanks again for your response, and I hope this email doesn't come across
as a sales pitch---we're just a couple of academics trying to create a
helpful learning resource :).
best wishes,
Colorado
On Sun, Nov 10, 2013 at 8:32 PM, Andy <t3k...@gmail.com> wrote:
> Hey Colorado and Roger.
> Thanks for your mail. Your project sounds very interesting.
> I haven't heard about it before, though. Is any one else on the list
> familiar with it?
> I have to concerns about adding it to the documentation:
>
> 1) I am not familiar with it and therefore have no way of judging the
> content. Maybe some other core devs do?
>
> 2) I am not sure how large the benefit would be. The documentation
> contains "see also" and "references" sections. You site basically seems
> to provide (very) advanced versions of these. From a user perspective,
> this seems to be an additional indirection.
> So if a user reads about a method, it seems easier to just give "Elements
> of statistical learning" as a reference instead of linking to
> metacademy, which then gives the reference.
>
> What do you think?
>
> Cheers,
> Andy
>
>
>
> On 11/08/2013 01:10 PM, colorado reed wrote:
>
> Hello,
>
> We're contributors to the open source project Metacademy (
> http://www.metacademy.org http://github.com/metacademy), which is
> creating a curated web of machine learning concepts. The idea is that
> Metacademy can tell users how to efficiently learn a concept and all of its
> necessary prerequisites given their current knowledge. One of the
> motivations for this project was to help practitioners with varying
> backgrounds learn the intricacies of the algorithms they use (without
> spending a few months taking an online course or a few thousand dollars
> taking a traditional one).
> Here's a quick example of a learning plan:
>
> http://metacademy.org/graphs/concepts/logistic_regression
>
> (click the graph button in the upper right to visualize the prerequsite
> structure)
> (try clicking a few check marks of the concepts you know and pressing
> "hide" in the toolbox in the upper right -- this creates a personalized
> learning plan)
>
> We thought adding pointers to metacademy concepts in the SKL
> documentation could help interested practitioners gain a deeper
> understanding of these concepts. But rather than changing the documentation
> willy-nilly and making pull requests, we thought we should contact the
> current contributors and see whether they feel this idea could benefit SKL.
> We currently have over 300 annotated concepts, which covers most of the
> functionality present in SKL, and we are happy to add any missing concepts
> that SKL users/contributors feel would benefit the documentation.
>
> best wishes
> Colorado Reed & Roger Grosse
>
>
> ------------------------------------------------------------------------------
> November Webinars for C, C++, Fortran Developers
> Accelerate application performance with scalable programming models. Explore
> techniques for threading, error checking, porting, and tuning. Get the most
> from the latest Intel processors and coprocessors. See abstracts and
> registerhttp://pubads.g.doubleclick.net/gampad/clk?id=60136231&iu=/4140/ostg.clktrk
>
>
>
> _______________________________________________
> Scikit-learn-general mailing
> listScikit-learn-general@lists.sourceforge.nethttps://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>
>
>
>
> ------------------------------------------------------------------------------
> November Webinars for C, C++, Fortran Developers
> Accelerate application performance with scalable programming models.
> Explore
> techniques for threading, error checking, porting, and tuning. Get the most
> from the latest Intel processors and coprocessors. See abstracts and
> register
> http://pubads.g.doubleclick.net/gampad/clk?id=60136231&iu=/4140/ostg.clktrk
> _______________________________________________
> Scikit-learn-general mailing list
> Scikit-learn-general@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>
>
------------------------------------------------------------------------------
DreamFactory - Open Source REST & JSON Services for HTML5 & Native Apps
OAuth, Users, Roles, SQL, NoSQL, BLOB Storage and External API Access
Free app hosting. Or install the open source package on any LAMP server.
Sign up and see examples for AngularJS, jQuery, Sencha Touch and Native!
http://pubads.g.doubleclick.net/gampad/clk?id=63469471&iu=/4140/ostg.clktrk
_______________________________________________
Scikit-learn-general mailing list
Scikit-learn-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
