Hi Gerard,
We're diverging from the initial thread. I'll respond to one point, we should take the rest of the discussion somewhere else. :)
On Mon, Apr 17, 2017 at 11:36 AM, Gerard Meijssen <gerard.meijssen@gmail.com
wrote:
So yes, your approach is good but like the translation tool it relies on English content.
No. The approach and implementation is from any language to any language. You can play with a very simplified version of the recommendations at https://recommend.wmflabs.org/. You can choose any language as source or destination.
Best, Leila
Thanks, GerardM
On 17 April 2017 at 18:40, Leila Zia leila@wikimedia.org wrote:
Hoi Gerard,
On Mon, Apr 17, 2017 at 7:54 AM, Gerard Meijssen < gerard.meijssen@gmail.com> wrote:
When you analyse articles and find that some things are missing, it
will
help a lot when you can target these articles to the people who are
likely
interested. When people interested in soccer learn that a soccer player died, they are more likely to edit even write an article.
You are absolutely right. This is what we even tested in the article creation recommendation experiment and you could see that providing personalized recommendations (where personalization was on the basis of matching editors interests based on their history of contributions) does better than random important recommendations. A few pointers for you:
- Check out section 2.3 of the paper at https://arxiv.org/pdf/1604.
03235.pdf to see how this was done.
- I talk briefly about how we do the editor interest modeling at
https://youtu.be/lHbnvRwFC_A?t=20m44s
In general, we have at least two ways for recommending to people what
they
like to edit: one would be using the information in their past edit
history
and building topical models that can help us learn what topics an editor
is
interested in. The other is by asking the editor to provide some seeds of interest to us. For example, we ask you to tell us what kind of article
you
would edit, and we give you recommendations similar to the seed you provide. Each have its own advantages and you sometimes have to mix the
two
approaches (and more) to give the editor enough breadth and depth of
topics
to choose from.
The approach for finding a subject that could do with more attention is
one
I applaud. When you want to do this across languages think Wikidata to define the area of interest for users. It will always include all the articles in all the languages. As you have seen with the Listeria
lists,
showing red links and Wikidata items is trivial.
Yes, finding what is missing in a Wikipedia language by comparing
language
editions is relatively easy, thanks to Wikidata. :) What is hard is
ranking
these millions of missing articles in any language based on some notion
of
importance. We developed a ranking system for the research I mentioned above. You can read about it in Section 2.2 of the paper at https://arxiv.org/pdf/1604.03235.pdf%E2%80%8B. I talk about in less details at https://youtu.be/lHbnvRwFC_A?t=16m58s. In a nutshell: we built a prediction model that aims to predict the number of pageviews the article would receive had it existed in the destination language where it's missing today. The higher this predicted number for a missing article in a language, the more important it is to create it.
Best, Leila
Thanks, Gerard
On 17 April 2017 at 02:04, Leila Zia leila@wikimedia.org wrote:
Hi John,
This may be of interest to you:
We are working on building recommendation systems than can help
editors
identify how to expand already existing articles in Wikipedia. This includes but is not limited to identifying what sections are missing
from
an article, what citations, what images, infobox information, etc.
This
is
research in its early days, if you'd like to follow up with it please
visit
Wikipedia_stubs_across_
languages
Best, Leila
Leila Zia Senior Research Scientist Wikimedia Foundation
On Sat, Apr 15, 2017 at 2:50 PM, John Erling Blad jeblad@gmail.com wrote:
Are anyone doing any work on automated quality assurance of
articles?
Not
the ORES-stuff, that is about creating hints from measured
features.
I'm
thinking about verifying existence and completeness of citations,
and
structure of logical arguments.
John _______________________________________________ Wikimedia-l mailing list, guidelines at:
wiki/Mailing_lists/Guidelines and https://meta.wikimedia.org/ wiki/Wikimedia-l New messages to: Wikimedia-l@lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/
mailman/listinfo/wikimedia-l,
<mailto:wikimedia-l-request@lists.wikimedia.org?subject=
unsubscribe>
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/ wiki/Mailing_lists/Guidelines and https://meta.wikimedia.org/ wiki/Wikimedia-l New messages to: Wikimedia-l@lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/
mailman/listinfo/wikimedia-l,
mailto:wikimedia-l-request@lists.wikimedia.org?subject=unsubscribe
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/ wiki/Mailing_lists/Guidelines and https://meta.wikimedia.org/ wiki/Wikimedia-l New messages to: Wikimedia-l@lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, mailto:wikimedia-l-request@lists.wikimedia.org?subject=unsubscribe
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/ wiki/Mailing_lists/Guidelines and https://meta.wikimedia.org/ wiki/Wikimedia-l New messages to: Wikimedia-l@lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, mailto:wikimedia-l-request@lists.wikimedia.org?subject=unsubscribe
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/ wiki/Mailing_lists/Guidelines and https://meta.wikimedia.org/ wiki/Wikimedia-l New messages to: Wikimedia-l@lists.wikimedia.org Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, mailto:wikimedia-l-request@lists.wikimedia.org?subject=unsubscribe