December 2017 Research Showcase

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

December 2017 Research Showcase

Lani Goto
Hi Everyone,

The next Research Showcase will be live-streamed this Wednesday, December
13, 2017 at 11:15 AM (PST) 18:15 UTC.

YouTube stream: https://www.youtube.com/watch?v=OoVwus1Owtk

As usual, you can join the conversation on IRC at #wikimedia-research. And,
you can watch our past research showcases here.

This month's presentation:
*The State of the Article Expansion Recommendation System*
By Leila Zia
Only 1% of English Wikipedia articles are labeled with quality class Good
or better, and 37% of the articles are stubs. We are building an article
expansion recommendation system to change this in Wikipedia, across many
languages. In this presentation, I will talk with you about our current
thinking of the vision and direction of the research that can help us build
such a recommendation system, and share more about one specific area of
research we have heavily focused on in the past months: building a
recommendation system that can help editors identify what sections to add
to an already existing article. I present some of the challenges we faced,
the methods we devised or used to overcome them, and the result of the
first line of experiments on the quality of such recommendations (teaser:
the results are really promising. The precision and recall at 10 is 80%.)


--
Lani Goto
Project Assistant, Engineering Admin
_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
Reply | Threaded
Open this post in threaded view
|

Re: December 2017 Research Showcase

James Salsman-2
Leila and Lani,

The Article Expansion Recommendation System is an absolutely
spectacular project, which will clearly very substantially improve the
encyclopedia in ways that perhaps no other single effort has come near
to being able, so I can't wait to learn more about it. But I might not
be able to make the live-stream time, so I want to get in this
question in advance:

Are you using or do you plan to use ORES quality predictions, the
upcoming article importance predictions, and pageview statistics to
rank article expansion recommendations?

> The next Research Showcase will be live-streamed this Wednesday, December
> 13, 2017 at 11:15 AM (PST) 18:15 UTC.
>
> YouTube stream: https://www.youtube.com/watch?v=OoVwus1Owtk
>
> As usual, you can join the conversation on IRC at #wikimedia-research. And,
> you can watch our past research showcases here.
>
> This month's presentation:
> "The State of the Article Expansion Recommendation System"
>
> By Leila Zia
>
> Only 1% of English Wikipedia articles are labeled with quality class Good
> or better, and 37% of the articles are stubs. We are building an article
> expansion recommendation system to change this in Wikipedia, across many
> languages. In this presentation, I will talk with you about our current
> thinking of the vision and direction of the research that can help us build
> such a recommendation system, and share more about one specific area of
> research we have heavily focused on in the past months: building a
> recommendation system that can help editors identify what sections to add
> to an already existing article. I present some of the challenges we faced,
> the methods we devised or used to overcome them, and the result of the
> first line of experiments on the quality of such recommendations (teaser:
> the results are really promising. The precision and recall at 10 is 80%.)

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
Reply | Threaded
Open this post in threaded view
|

Re: December 2017 Research Showcase

Leila Zia
Hi James,

On Tue, Dec 12, 2017 at 5:57 AM, James Salsman <[hidden email]> wrote:
> Leila and Lani,
>
> The Article Expansion Recommendation System is an absolutely
> spectacular project, which will clearly very substantially improve the
> encyclopedia in ways that perhaps no other single effort has come near
> to being able, so I can't wait to learn more about it.

Thanks for the words of encouragement. :)

> But I might not
> be able to make the live-stream time,

A quick note (which you already probably know): you can watch the
showcase afterwards, and you know where to find us, so do ping with
questions. (I may be slow in responding in the next 2 weeks, but I
will get back to you).

> so I want to get in this
> question in advance:
>
> Are you using or do you plan to use ORES quality predictions, the
> upcoming article importance predictions, and pageview statistics to
> rank article expansion recommendations?

At the moment, we're not using the systems you mentioned but that's
because we're not at the ranking stage, yet. Once we get there, we
will start interacting with them more, to figure out where we can use
what. In our previous work for recommending what articles to be
created, we ranked missing articles in a given language (Section 2.2.
in https://arxiv.org/abs/1604.03235). When we're working with already
existing articles, our job can become easier, given that, as you
mentioned, more information is available about the article. (For
example, we had to build a prediction model that would predict article
pageviews if they get created in the previous work, while pageviews
are now available to us for the already existing articles.).

While we're at it: While ranking is an important step, so far what we
have learned is that there is so much missing articles/content across
languages that it suffices if the ranking is "good enough". The
combination of a good enough ranking to give us, let's say the top
300K thousand high priority items to contribute to, and empowering
users to tune recommendations based on their interests will be
powerful enough from the users' perspective.

I will add a few slides to talk about the challenges we see if article
expansion, inspired by your question. :)

Best,
Leila

>
>> The next Research Showcase will be live-streamed this Wednesday, December
>> 13, 2017 at 11:15 AM (PST) 18:15 UTC.
>>
>> YouTube stream: https://www.youtube.com/watch?v=OoVwus1Owtk
>>
>> As usual, you can join the conversation on IRC at #wikimedia-research. And,
>> you can watch our past research showcases here.
>>
>> This month's presentation:
>> "The State of the Article Expansion Recommendation System"
>>
>> By Leila Zia
>>
>> Only 1% of English Wikipedia articles are labeled with quality class Good
>> or better, and 37% of the articles are stubs. We are building an article
>> expansion recommendation system to change this in Wikipedia, across many
>> languages. In this presentation, I will talk with you about our current
>> thinking of the vision and direction of the research that can help us build
>> such a recommendation system, and share more about one specific area of
>> research we have heavily focused on in the past months: building a
>> recommendation system that can help editors identify what sections to add
>> to an already existing article. I present some of the challenges we faced,
>> the methods we devised or used to overcome them, and the result of the
>> first line of experiments on the quality of such recommendations (teaser:
>> the results are really promising. The precision and recall at 10 is 80%.)

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
Reply | Threaded
Open this post in threaded view
|

Re: December 2017 Research Showcase

Lani Goto
In reply to this post by Lani Goto
Hi everyone, the showcase will be starting in 30 minutes.

On Mon, Dec 11, 2017 at 5:22 PM, Lani Goto <[hidden email]> wrote:

> Apologies, a clarification about the time:
>
> This will be at 11:30AM (PST) 19:30 UTC.
>
> On Mon, Dec 11, 2017 at 4:21 PM, Lani Goto <[hidden email]> wrote:
>
>> Hi Everyone,
>>
>> The next Research Showcase will be live-streamed this Wednesday, December
>> 13, 2017 at 11:15 AM (PST) 18:15 UTC.
>>
>> YouTube stream: https://www.youtube.com/watch?v=OoVwus1Owtk
>>
>> As usual, you can join the conversation on IRC at #wikimedia-research.
>> And, you can watch our past research showcases here.
>>
>> This month's presentation:
>> *The State of the Article Expansion Recommendation System*
>> By Leila Zia
>> Only 1% of English Wikipedia articles are labeled with quality class Good
>> or better, and 37% of the articles are stubs. We are building an article
>> expansion recommendation system to change this in Wikipedia, across many
>> languages. In this presentation, I will talk with you about our current
>> thinking of the vision and direction of the research that can help us build
>> such a recommendation system, and share more about one specific area of
>> research we have heavily focused on in the past months: building a
>> recommendation system that can help editors identify what sections to add
>> to an already existing article. I present some of the challenges we faced,
>> the methods we devised or used to overcome them, and the result of the
>> first line of experiments on the quality of such recommendations (teaser:
>> the results are really promising. The precision and recall at 10 is 80%.)
>>
>>
>> --
>> Lani Goto
>> Project Assistant, Engineering Admin
>>
>
>
>
> --
> Lani Goto
> Project Assistant, Engineering Admin
>



--
Lani Goto
Project Assistant, Engineering Admin
_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l