This list has been dormant for several years now - all traffic has
been multiply-cc'ed announcements.
As such, I'm going to recommend that with the mailman3 upgrade (which
is finally happening!), we shut and archive this list. Past posts
should still be available - this list is important historically, but
hasn't been a current communications channel for years.
- d.
Forwarding.
Pine
( https://meta.wikimedia.org/wiki/User:Pine )
---------- Forwarded message ---------
From: Janna Layton <jlayton(a)wikimedia.org>
Date: Fri, May 15, 2020 at 8:05 PM
Subject: [Wiki-research-l] [Wikimedia Research Showcase] May 20, 2020:
Human in the Loop Machine Learning
To: <analytics(a)lists.wikimedia.org>,
<wikimedia-l(a)lists.wikimedia.org>,
<wiki-research-l(a)lists.wikimedia.org>
Hi all,
The next Research Showcase will be live-streamed on Wednesday, May 20, at
9:30 AM PDT/16:30 UTC.
This month we will learn about recent research on machine learning systems
that rely on human supervision for their learning and optimization -- a
research area commonly referred to as Human-in-the-Loop ML. In the first
talk, Jie Yang will present a computational framework that relies on
crowdsourcing to identify influencers in Social Networks (Twitter) by
selectively obtaining labeled data. In the second talk, Estelle Smith will
discuss the role of the community in maintaining ORES, the machine learning
system that predicts the quality in Wikipedia applications.
YouTube stream: https://www.youtube.com/watch?v=8nDiu2ebdOI
As usual, you can join the conversation on IRC at #wikimedia-research. You
can also watch our past research showcases here:
https://www.mediawiki.org/wiki/Wikimedia_Research/Showcase
This month's presentations:
*OpenCrowd: A Human-AI Collaborative Approach for Finding Social
Influencers via Open-Ended Answers Aggregation*
By: Jie Yang, Amazon (current), Delft University of Technology (starting
soon)
Finding social influencers is a fundamental task in many online
applications ranging from brand marketing to opinion mining. Existing
methods heavily rely on the availability of expert labels, whose collection
is usually a laborious process even for domain experts. Using open-ended
questions, crowdsourcing provides a cost-effective way to find a large
number of social influencers in a short time. Individual crowd workers,
however, only possess fragmented knowledge that is often of low quality. To
tackle those issues, we present OpenCrowd, a unified Bayesian framework
that seamlessly incorporates machine learning and crowdsourcing for
effectively finding social influencers. To infer a set of influencers,
OpenCrowd bootstraps the learning process using a small number of expert
labels and then jointly learns a feature-based answer quality model and the
reliability of the workers. Model parameters and worker reliability are
updated iteratively, allowing their learning processes to benefit from each
other until an agreement on the quality of the answers is reached. We
derive a principled optimization algorithm based on variational inference
with efficient updating rules for learning OpenCrowd parameters.
Experimental results on finding social influencers in different domains
show that our approach substantially improves the state of the art by 11.5%
AUC. Moreover, we empirically show that our approach is particularly useful
in finding micro-influencers, who are very directly engaged with smaller
audiences.
Paper: https://dl.acm.org/doi/fullHtml/10.1145/3366423.3380254
*Keeping Community in the Machine-Learning Loop*
By: C. Estelle Smith, MS, PhD Candidate, GroupLens Research Lab at the
University of Minnesota
On Wikipedia, sophisticated algorithmic tools are used to assess the
quality of edits and take corrective actions. However, algorithms can fail
to solve the problems they were designed for if they conflict with the
values of communities who use them. In this study, we take a
Value-Sensitive Algorithm Design approach to understanding a
community-created and -maintained machine learning-based algorithm called
the Objective Revision Evaluation System (ORES)—a quality prediction system
used in numerous Wikipedia applications and contexts. Five major values
converged across stakeholder groups that ORES (and its dependent
applications) should: (1) reduce the effort of community maintenance, (2)
maintain human judgement as the final authority, (3) support differing
peoples’ differing workflows, (4) encourage positive engagement with
diverse editor groups, and (5) establish trustworthiness of people and
algorithms within the community. We reveal tensions between these values
and discuss implications for future research to improve algorithms like
ORES.
Paper:
https://commons.wikimedia.org/wiki/File:Keeping_Community_in_the_Loop-_Unde…
--
Janna Layton (she, her)
Administrative Assistant - Product & Technology
Wikimedia Foundation <https://wikimediafoundation.org/>
_______________________________________________
Wiki-research-l mailing list
Wiki-research-l(a)lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
Dear colleagues,
The English Wikipedia community received news that User:Brianboulton
<https://en.wikipedia.org/wiki/User:Brianboulton> passed. Quoting
User:Schrocat <https://en.wikipedia.org/wiki/User:SchroCat>: "I have just
received an email from Brian's daughter to say that he died peacefully on 9
December, following a long illness. Requiescat in pace. 106 FAs, 2 FLs,
gawd knows how many source and prose reviews at FA, and countless numbers
of editors helped, encouraged and improved over the years. A good friend to
all who met him, and this place is a little less appealing now he won't be
here anymore."
Comments may be left on his talk page <https://en.wikipedia.org/wiki/User
talk:Brianboulton>.
Regards,
Pine
( https://meta.wikimedia.org/wiki/User:Pine )
Hi Aron,
Thanks for sharing your thoughts. However, I think that the more
appropriate mailing list for this is WikiEN-l. I'm forwarding this thread
there.
Regarding your second point, I think that many admins are not selfish
people and are often willing to criticize other admins, so I am not
concerned that admins are participating in this discussion. I would be
concerned if I felt that admins as a group are power hungry and are more
interested in defending each other than in working for the best interests
of the encyclopedia, but that is not my impression. I think that many
admins are highly skilled, generous with their time, and willing to
volunteer for thankless and sometimes personally risky activities. Those
admins that cause problems are, sooner or later, usually in trouble with
the Arbitration Committee. That is not to say that I'm opposed to a new
desysop procedure, but I believe that most English Wikipedia admins are
competent and act in good faith.
Pine
( https://meta.wikimedia.org/wiki/User:Pine )
---------- Forwarded message ---------
From: Aron Manning <aronmanning5(a)gmail.com>
Date: Sat, Oct 19, 2019 at 10:06 AM
Subject: [Wikimedia-l] RfC: binding desysop procedure on English Wikipedia
To: Wikimedia Mailing List <wikimedia-l(a)lists.wikimedia.org>
https://en.wikipedia.org/wiki/Wikipedia:Requests_for_comment/2019_community…
So far about 60% of the commenters are administrators, which role comes
with an inherent and unavoidable conflict-of-interest. Although the
discussion is quite neutral, administrators are a minor part of the
community, thus hopefully editors will join the discussion as well, and
share their views and suggestions on the topic.
Aron
_______________________________________________
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(a)lists.wikimedia.org
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
<mailto:wikimedia-l-request@lists.wikimedia.org?subject=unsubscribe>
Early warning of planned technical maintenance window for ENWP.
Pine
( https://meta.wikimedia.org/wiki/User:Pine )
---------- Forwarded message ---------
From: Manuel Arostegui <marostegui(a)wikimedia.org>
Date: Mon, Oct 7, 2019 at 9:20 AM
Subject: [Wikitech-l] s1 (enwiki) primary master switchover (read-only
required) 14th Nov 05:00 AM UTC
To: Operations Engineers <ops(a)lists.wikimedia.org>, Wikimedia developers <
wikitech-l(a)lists.wikimedia.org>
Hello,
We have requested a 30 minutes read-only window for s1 (enwiki) (T234801)
for the 14th November from 05:00-05:30 AM UTC to switchover that section
primary database master (T234800)
db1067 is an old host and out of warranty that will be decommissioned
(T217396). The new master will be db1083
We are going to do this on Thursday 14th Nov from 05:00 to 05:30 AM UTC (we
do not expect to use the 30 minutes window, if everything goes as expected).
Impact: Writes will be blocked on the following wiki:
enwiki
Reads will remain unaffected.
Communication will happen at #wikimedia-operations
If you are around at that time and want to help with the monitoring, please
join us!
Thanks
_______________________________________________
Wikitech-l mailing list
Wikitech-l(a)lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Forwarding in case this is of interest to others.
Pine
( https://meta.wikimedia.org/wiki/User:Pine )
---------- Forwarded message ---------
From: Janna Layton <jlayton(a)wikimedia.org>
Date: Thu, Oct 10, 2019 at 9:58 PM
Subject: [Analytics] [Wikimedia Research Showcase] October 16, 2019 at 9:30
AM PDT, 16:30 UTC
To: <wikimedia-l(a)lists.wikimedia.org>, <analytics(a)lists.wikimedia.org>, <
wiki-research-l(a)lists.wikimedia.org>
Hi all,
The next Research Showcase will be live-streamed next Wednesday, October
16, at 9:30 AM PDT/16:30 UTC.
YouTube stream: https://www.youtube.com/watch?v=KZ35weAVlIU
As usual, you can join the conversation on IRC at #wikimedia-research. You
can also watch our past Research Showcases here:
https://www.mediawiki.org/wiki/Wikimedia_Research/Showcase
This month's presentations:
Elections Without Fake: Deploying Real Systems to Counter Misinformation
Campaigns
By Fabrício Benevenuto, Computer Science Department, Universidade Federal
de Minas Gerais (UFMG), Brazil
The political debate and electoral dispute in the online space during the
2018 Brazilian elections were marked by an information war. In order to
mitigate the misinformation problem, we created the project Elections
Without Fake <http://www.eleicoes-sem-fake.dcc.ufmg.br/> and developed a
few technological solutions able to reduce the abuse of misinformation
campaigns in the online space. Particularly, we created a system to monitor
public groups in WhatsApp and a system to monitor ads in Facebook. Our
systems showed to be fundamental for fact-checking and investigative
journalism, and are currently being used by over 150 journalists with
editorial lines and various fact-checking agencies.
More info on second talk by Francesca Spezzano to come
--
Janna Layton (she, her)
Administrative Assistant - Product & Technology
Wikimedia Foundation <https://wikimediafoundation.org/>
_______________________________________________
Analytics mailing list
Analytics(a)lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/analytics
Forwarding an announcement.
(Personal comments: what might appear to be consistent application of
policies to one person might appear to be bullying to someone else. If I
find the time to watch this video, I will be interested to hear from the
researchers regarding this issue. I think that both of the presentations
sound interesting.)
Regards,
Pine
( https://meta.wikimedia.org/wiki/User:Pine )
---------- Forwarded message ---------
From: Janna Layton <jlayton(a)wikimedia.org>
Date: Thu, Jul 11, 2019 at 11:46 PM
Subject: [Analytics] [Wikimedia Research Showcase] July 17, 2019 at 11:30
AM PDT, 18:30 UTC
To: <wikimedia-l(a)lists.wikimedia.org>, <analytics(a)lists.wikimedia.org>, <
wiki-research-l(a)lists.wikimedia.org>
Hi all,
The next Research Showcase will be live-streamed next Wednesday, July 17,
at 11:30 AM PDT/18:30 UTC.
YouTube stream: https://www.youtube.com/watch?v=i9vvwV5KfW4
As usual, you can join the conversation on IRC at #wikimedia-research. You
can also watch our past research showcases here:
https://www.mediawiki.org/wiki/Wikimedia_Research/Showcase
This month's presentations:
Characterizing Incivility on Wikipedia
Elizabeth Whittaker, University of Michigan School of Information
In a society whose citizens have a variety of viewpoints, there is a
question of how citizens can govern themselves in ways that allow these
viewpoints to co-exist. Online deliberation has been posited as a problem
solving mechanism in this context, and civility can be thought of as a
mechanism that facilitates this deliberation. Civility can thus be thought
of as a method of interaction that encourages collaboration, while
incivility disrupts collaboration. However, it is important to note that
the nature of online civility is shaped by its history and the technical
architecture scaffolding it. Civility as a concept has been used both to
promote equal deliberation and to exclude the marginalized from
deliberation, so we should be careful to ensure that our conceptualizations
of incivility reflect what we intend them to in order to avoid
unintentionally reinforcing inequality.
To this end, we examined Wikipedia editors’ perceptions of interactions
that disrupt collaboration through 15 semi-structured interviews. Wikipedia
is a highly deliberative platform, as editors need to reach consensus about
what will appear on the article page, a process that often involves
deliberation to coordinate, and any disruption to this process should be
apparent. We found that incivility on Wikipedia typically occurs in one of
three ways: through weaponization of Wikipedia’s policies, weaponization of
Wikipedia’s technical features, and through more typical vitriolic content.
These methods of incivility were gendered, and had the practical effect of
discouraging women from editing. We implicate this pattern as one of the
underlying causes of Wikipedia’s gender gap.
Hidden Gems in the Wikipedia Discussions: The Wikipedians’ Rationales
Lu Xiao, Syracuse University School of Information Studies
I will present a series of completed and ongoing studies that are aimed at
understanding the role of the Wikipedians’ rationales in Wikipedia
discussions. We define a rationale as one’s justification of her viewpoint
and suggestions. Our studies demonstrate the potential of leveraging the
Wikipedians’ rationales in discussions as resources for future
decision-making and as resources for eliciting knowledge about the
community’s norms, practices and policies. Viewed as rich digital traces in
these environments, we consider them to be beneficial for the community
members, such as helping newcomers familiarize themselves on the commonly
accepted justificatory reasoning styles. We call for more research
attention to the discussion content from this rationale study perspective.
--
Janna Layton (she, her)
Administrative Assistant - Audiences & Technology
Wikimedia Foundation <https://wikimediafoundation.org/>
_______________________________________________
Analytics mailing list
Analytics(a)lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/analytics