Return to Video

cdn.media.ccc.de/.../wikidatacon2019-1147-eng-Project_WikiLoop_hd.mp4

  • 0:07 - 0:12
    [inaudible] and I
    have an effort called WikiLoop,
  • 0:12 - 0:15
    and this is what I'm going
    to introduce to you about.
  • 0:16 - 0:23
    We have presented WikiLoop, the idea,
    to several Wikimedia related conferences.
  • 0:23 - 0:25
    How many of you have heard
    about WikiLoop before?
  • 0:26 - 0:27
    Thanks.
  • 0:27 - 0:31
    And how many of you have interacted
    with the datasets and toolings
  • 0:31 - 0:33
    that we provided before?
  • 0:33 - 0:37
    Okay, fairly new.
    So this will be mostly an introduction.
  • 0:37 - 0:42
    So we would like to tell you
    why we start this initiative
  • 0:42 - 0:44
    and what it intends to do,
  • 0:44 - 0:49
    and how you can get involved
    or what it will go for.
  • 0:50 - 0:54
    So, to begin with,
    we would like to give you an example.
  • 0:54 - 0:58
    This is a vandalism
    that happened in Italian...
  • 1:01 - 1:04
    that happened in Italy Wikipedia.
  • 1:04 - 1:07
    I know that most people here
    are interested in Wikidata.
  • 1:07 - 1:10
    I will tell you why this is relevant too.
  • 1:10 - 1:12
    So basically what we found is
  • 1:12 - 1:16
    that someone vandalized
    Wikipedia on Italian
  • 1:16 - 1:21
    and says, "Bezos who cannot afford a car."
  • 1:21 - 1:23
    And this is an interesting question,
  • 1:24 - 1:28
    if you think about it,
    this is blatant obvious vandalism
  • 1:28 - 1:33
    but when it comes to machines
    and algorithms
  • 1:33 - 1:38
    which find to detect vandalism
    and avoid serving users the information,
  • 1:38 - 1:42
    how can computer understand
    this kind of information,
  • 1:42 - 1:43
    like it would be...
  • 1:47 - 1:49
    we realize that sometimes
    there are limitations
  • 1:49 - 1:54
    of how far algorithms can go
    and machine can go.
  • 1:55 - 1:58
    Another example here is let's say,
  • 1:58 - 2:02
    there is a word or label,
    or a category on Wikipedia says,
  • 2:02 - 2:06
    someone, a person,
    is a Christian scientist.
  • 2:06 - 2:10
    Now, given this label,
    what facts do you come up with
  • 2:10 - 2:14
    like what would you infer
    from this category?
  • 2:14 - 2:19
    Do you think it would be a "Christian"
    or do you think it would be a "scientist"?
  • 2:19 - 2:22
    In this specific case--
    it does not apply everywhere--
  • 2:22 - 2:23
    but it this specific case,
  • 2:23 - 2:27
    there is a religion
    called "Christian Science,"
  • 2:27 - 2:30
    and people who hold that belief
    is called "Christian Scientist."
  • 2:32 - 2:35
    And, again, for machines,
    how can we know, like
  • 2:36 - 2:40
    even if many people here are big [fan]
  • 2:40 - 2:45
    that's the better we make our data
    a knowledge machine-friendly
  • 2:45 - 2:52
    the easier we can work and improve
    the overall knowledge accessibility
  • 2:52 - 2:54
    and contribute together
  • 2:54 - 2:56
    but there is always things
  • 2:56 - 2:58
    that we believe
    that machine has restrictions.
  • 3:00 - 3:04
    So all in all, we start to realize
  • 3:04 - 3:08
    that coming from Internet companies
  • 3:08 - 3:11
    who have a strong belief
    of our technology
  • 3:11 - 3:13
    and what machine can do,
  • 3:13 - 3:16
    there is always a gap
    or there is always something
  • 3:16 - 3:19
    that we would need to rely on human being
  • 3:19 - 3:22
    and more, we would need
    to rely on communities
  • 3:23 - 3:28
    who are actively contributing,
    who are doing the peer reviews to our...
  • 3:28 - 3:30
    collaborating with each other.
  • 3:30 - 3:36
    So this is a picture
    about the background effort of WikiLoop.
  • 3:37 - 3:40
    For the human being,
    they have the knowledge,
  • 3:40 - 3:46
    we have our domain expertize
    and we can crosscheck each other
  • 3:46 - 3:49
    but we just have that enough time.
  • 3:49 - 3:53
    And there are many things
    that machine can empower this
  • 3:53 - 3:56
    but there is restrictions there.
  • 3:56 - 3:59
    So the goal is to empower
  • 3:59 - 4:03
    or improve the productivity
    of human editors.
  • 4:03 - 4:09
    But also the other side of the formula
    is we want to loop that back
  • 4:09 - 4:13
    to the research and the academic efforts
  • 4:13 - 4:17
    that improve how machine
    can help in these cases.
  • 4:18 - 4:23
    So by raise of hand,
    how many of you have used Google?
  • 4:24 - 4:25
    Thank you.
  • 4:25 - 4:26
    And how many of you
  • 4:27 - 4:31
    think that companies like Google
    and other big knowledge companies
  • 4:31 - 4:34
    should contribute more
    to the knowledge world?
  • 4:36 - 4:38
    So what happens is that...
  • 4:38 - 4:42
    we all know that our mission at Google
    or other similar companies--
  • 4:42 - 4:48
    we have a strong background
    of leveraging the open knowledge world,
  • 4:48 - 4:50
    like for Google specific case
  • 4:50 - 4:53
    it's like organize
    the world's information.
  • 4:53 - 4:55
    So we help disseminate the information,
  • 4:56 - 5:00
    which in one sense that helps
    the mission of this movement.
  • 5:00 - 5:06
    But only every once a while
    we have sporadic help
  • 5:08 - 5:12
    trying to donate knowledge
    and datasets, and tools,
  • 5:12 - 5:16
    and we want to see
    if we can make this sustainable,
  • 5:18 - 5:21
    both in the technical sense
  • 5:21 - 5:23
    and also in the business sense.
  • 5:25 - 5:30
    So this is like
    a one-sentence introduction.
  • 5:30 - 5:35
    We want WikiLoop
    to become an umbrella program
  • 5:35 - 5:37
    for a series of technical projects
  • 5:37 - 5:40
    intended to contribute
    datasets and toolings
  • 5:40 - 5:45
    and hopefully make this a community effort
    with participation of
  • 5:45 - 5:50
    other likeminded people,
    partners and institutions
  • 5:50 - 5:52
    to join with this effort.
  • 5:52 - 5:56
    There are several projects
    that we think would be a good fit,
  • 5:56 - 5:59
    and these are the criteria.
  • 5:59 - 6:04
    First of all, the idea is
    that it needs to be source improvements
  • 6:04 - 6:07
    or source improvements
    by and large is a good fit,
  • 6:07 - 6:11
    and also the second thing
    that companies like us
  • 6:11 - 6:14
    really cannot do very well by ourself
  • 6:14 - 6:18
    is to maximize the neutrality,
    to avoid picking sides
  • 6:18 - 6:22
    on the controversies,
    decisions or discussions
  • 6:22 - 6:27
    and another thing is that to make this
    in the long-term sustainability
  • 6:27 - 6:32
    and to keep it
    being supported by this industry.
  • 6:32 - 6:35
    We want to see
    the productivity, the scalability
  • 6:35 - 6:38
    of our contribution and efforts.
  • 6:38 - 6:41
    To explain a little bit more...
  • 6:42 - 6:44
    We always look trying to extract...
  • 6:44 - 6:47
    for example, we are trying
    to extract facts from Wikipedia.
  • 6:47 - 6:53
    And while we can do
    several separations,
  • 6:53 - 6:56
    we're labeling, fairly well,
  • 6:56 - 7:00
    up to certain point
    the bottleneck is no longer
  • 7:00 - 7:02
    how good the machine,
    the algorithm can reach
  • 7:02 - 7:06
    but sometimes
    there is a noise in the source,
  • 7:06 - 7:11
    and if we do not remove the source
  • 7:11 - 7:14
    or minimize the source noise there,
  • 7:14 - 7:16
    that's how far the machine can go.
  • 7:16 - 7:18
    So that's the first criteria.
  • 7:18 - 7:19
    And the second criteria is,
  • 7:19 - 7:24
    we don't want to get to be seen as buyers
    or introduce potential buyers.
  • 7:24 - 7:30
    We want to rely on
    governance that is peer reviewed
  • 7:30 - 7:33
    and that is done by the community
  • 7:33 - 7:37
    so that we can avoid picking sides
    in the controversy questions.
  • 7:37 - 7:41
    And the third thing
    which probably not so intuitive
  • 7:41 - 7:43
    but this is the kind of...
    I would like...
  • 7:43 - 7:48
    Let me give you an example
    of the projects we have in mind.
  • 7:48 - 7:52
    Let's say there are smaller,
    minority language there.
  • 7:52 - 7:56
    I have heard a very good talk
    earlier this morning.
  • 7:56 - 7:58
    But one idea we have here is,
  • 7:58 - 8:02
    let's say you are a minority
    language contributor, very active,
  • 8:02 - 8:07
    and you want to advocate for your culture
    and supporting your knowledge creation.
  • 8:08 - 8:12
    But because companies like Google
    or other consumer company,
  • 8:12 - 8:15
    they have a bar
    for releasing a translation,
  • 8:15 - 8:16
    to make it available.
  • 8:16 - 8:19
    They want the precision to be high enough
  • 8:19 - 8:22
    so that they can use it to serve users.
  • 8:23 - 8:27
    But maybe internally they have AI modules
    that are experimenting,
  • 8:27 - 8:29
    not good enough to the bar
  • 8:29 - 8:31
    because lack of training data,
  • 8:33 - 8:35
    so the translation is not available.
  • 8:35 - 8:38
    But the community is doing
    the translation by hand anyway.
  • 8:39 - 8:41
    Now, one of the things we are thinking of,
  • 8:41 - 8:45
    if we can provide
    some of this experimental thing
  • 8:45 - 8:48
    that is not good enough
    to serve general user purpose
  • 8:48 - 8:50
    but still good for the community
  • 8:50 - 8:54
    and somewhat improving the productivity,
  • 8:54 - 8:56
    it would be able to
  • 8:56 - 9:01
    one, improve the speed of how well
    a community can contribute,
  • 9:01 - 9:06
    and second, what a community is creating
    anyway can come back as a training data
  • 9:06 - 9:09
    that keeps bootstrapping the machines.
  • 9:10 - 9:15
    So over time by this effort
    we hope to generate a model
  • 9:16 - 9:19
    that both helps
    the human being, the editors,
  • 9:19 - 9:22
    but also helps the research
  • 9:22 - 9:27
    that improves the AI and other approaches.
  • 9:28 - 9:32
    And this is a big overview
    of a few projects
  • 9:32 - 9:34
    we are going to introduce.
  • 9:34 - 9:37
    Due to the time limitation
    I will feature a few.
  • 9:37 - 9:41
    The WikiLoop Game, which you can look up,
  • 9:41 - 9:47
    is one that we leveraged a platform
  • 9:47 - 9:50
    created by Magnus called Wikidata Game.
  • 9:50 - 9:55
    We provide several datasets there
    to be played, to be introduced
  • 9:55 - 9:57
    and commit to the Wikidata
  • 9:57 - 9:59
    but by the human review.
  • 10:00 - 10:04
    And Google doesn't get
    to contribute data directly
  • 10:04 - 10:06
    to Wikipedia or Wikidata
  • 10:06 - 10:12
    but having someone who is reviewing it
    as non-biased individuals to do so.
  • 10:13 - 10:17
    And the second one I'm going to feature
    is WikiLoop Battlefield,
  • 10:17 - 10:21
    the one that you have seen just now
    as a counter-vandalism platform,
  • 10:21 - 10:26
    and this one also features
    the same criteria
  • 10:26 - 10:28
    of source improvements,
  • 10:30 - 10:33
    of how it can empower machines
  • 10:33 - 10:39
    by looping back to the training data
  • 10:39 - 10:43
    and also how it avoids companies like us
  • 10:43 - 10:49
    to pick sides allowing way to rely
    on the community's assessment.
  • 10:49 - 10:54
    And the third one is CitePool,
    which is creating...
  • 10:54 - 10:58
    we're trying to help creating
    citation candidate pool
  • 10:58 - 11:03
    to improve the productivity of people
    who want to add citation
  • 11:03 - 11:05
    but also see if we can make that
  • 11:05 - 11:10
    into a training data
    accessible to researchers.
  • 11:10 - 11:13
    So let me use WikiLoop Battlefield
    as an example.
  • 11:13 - 11:18
    If you have... try it on your phone--
    battlefield.wikiloop.org.
  • 11:18 - 11:22
    By the way, I want to highlight,
    the name is subject to change
  • 11:22 - 11:26
    because some friendly community members
    have come to me and suggest
  • 11:26 - 11:32
    that Battlefield might not be
    the best name for a project
  • 11:32 - 11:35
    serving the Wikimedia movement.
  • 11:35 - 11:40
    So if you don't like this name,
    come join us in the discussion,
  • 11:40 - 11:41
    provide your suggestion,
  • 11:41 - 11:44
    we will be very happy
    to converge to a name
  • 11:44 - 11:48
    that has community consensus
    and popularity.
  • 11:48 - 11:51
    But let's use that as a placeholder here.
  • 11:53 - 11:56
    I don't need to introduce
    to this group of people
  • 11:56 - 11:59
    about the typical vandalism workflow
  • 12:00 - 12:03
    but if you have already...
  • 12:05 - 12:09
    trying to conduct
    some counter-vandalism activity,
  • 12:09 - 12:12
    you might know that it's not very trivial.
  • 12:12 - 12:16
    How many of you have seen vandalism
    on Wikipedia and Wikidata?
  • 12:17 - 12:22
    Okay, how many of you
    have reverted, by hand, some of them?
  • 12:23 - 12:28
    How many of you have used certain tools
    or go ahead and find certain tools
  • 12:28 - 12:31
    to patrol or revert vandalism?
  • 12:31 - 12:32
    Okay.
  • 12:33 - 12:36
    Cool, this is
    the highest density of people
  • 12:36 - 12:41
    who have tried to revert vandalism
  • 12:41 - 12:44
    that I have spoken to before.
  • 12:44 - 12:49
    So maybe some of you have been
    very comfortably doing that
  • 12:49 - 12:53
    but for me as someone
    who started editing actively
  • 12:54 - 12:57
    only since like three years ago
  • 12:58 - 13:03
    and who only started to be very serous
    doing vandalism detection and patrolling
  • 13:04 - 13:06
    only since about last year
  • 13:06 - 13:11
    I found that doing so is not super easy
  • 13:11 - 13:14
    on the world of Wikimedia movement.
  • 13:15 - 13:22
    If we look at the existing alternatives
  • 13:22 - 13:26
    there are tools that is built
    featuring desktops,
  • 13:26 - 13:31
    there are tools that is relying
    on users who have rollback permissions,
  • 13:31 - 13:34
    which itself is a big barrier to get.
  • 13:35 - 13:39
    We want to make this
    a super easy to use platform
  • 13:39 - 13:42
    for all the three roles.
  • 13:42 - 13:46
    The first one is user, reviewer or editor,
    whatever you call it.
  • 13:47 - 13:48
    The second one is researcher
  • 13:48 - 13:53
    who is trying to create
    vandalism detecting algorithms or systems.
  • 13:53 - 13:55
    And the third one is developers
  • 13:55 - 14:00
    who is trying to improve
    this WikiLoop Battlefield tooling.
  • 14:00 - 14:02
    We want it to be
    super easy for user to use.
  • 14:02 - 14:05
    You can you pull up your phone,
    you don't have to install it,
  • 14:05 - 14:07
    you can do in on your laptop.
  • 14:07 - 14:10
    And we also want
    to lower a barrier to review.
  • 14:10 - 14:17
    The reason why other tools
    are trying to limit the access to the tool
  • 14:17 - 14:22
    is because there needs to be
    a base trust level for people to use them.
  • 14:22 - 14:27
    You don't want someone
    to come to a counter-vandalism tool
  • 14:27 - 14:28
    to vandalize itself.
  • 14:29 - 14:32
    So what we are trying to do is that,
  • 14:32 - 14:34
    to begin with, we want
    to make it super easy
  • 14:34 - 14:40
    but also we want to allow multiple people
    to label the same thing.
  • 14:40 - 14:42
    Also we want to make it super convenient
  • 14:42 - 14:48
    to see the [inaudible],
    to see other label, and all in real time.
  • 14:48 - 14:52
    We also want to make it
    for researchers super easy to use.
  • 14:52 - 14:55
    By one click you can download the labeling
  • 14:55 - 15:01
    and maybe start play with the data
    and see how it fits in your model.
  • 15:02 - 15:06
    And we provide APIs
    that have access to real time data.
  • 15:07 - 15:10
    And for the developer
    we make it very easy to pick up--
  • 15:10 - 15:15
    we have one click--
    you can deploy your trial instances,
  • 15:15 - 15:17
    things like that.
  • 15:17 - 15:21
    This is an example
    about building projects
  • 15:21 - 15:23
    for umbrella like WikiLoop.
  • 15:23 - 15:28
    We want to make sure
    the community trust comes the first.
  • 15:28 - 15:31
    We usually need to make it
    open source the best.
  • 15:32 - 15:37
    And we want to avoid proprietary tech,
    we want to avoid tech lock-down,
  • 15:38 - 15:43
    and we rely on community approval
    for certain features.
  • 15:44 - 15:49
    And if you have seen this--
    this is the components that we rely on--
  • 15:49 - 15:56
    still very early stage but you get
    the principles behind the design.
  • 15:56 - 16:00
    So what's next, we are trying
    to grow our usage.
  • 16:00 - 16:02
    Hopefully you can try it out by yourself
  • 16:02 - 16:07
    and promise to me
    that you don't click on the login.
  • 16:08 - 16:09
    There is a login button--
  • 16:09 - 16:10
    there will be some good features
  • 16:10 - 16:13
    that make it super easy
    to even revert something.
  • 16:13 - 16:15
    Currently it's still a jump to revert.
  • 16:17 - 16:18
    But we are building features,
  • 16:18 - 16:24
    and we are also trying
    to let you choose some categories
  • 16:24 - 16:27
    or the watchlist
    that you will be watching
  • 16:27 - 16:31
    and the one that you care about to patrol.
  • 16:32 - 16:38
    And also if you are researchers
    while doing related vandalism detection,
  • 16:38 - 16:42
    try our data and give us feedback.
  • 16:44 - 16:47
    And I will go through quickly
    about a few other projects
  • 16:47 - 16:49
    that we are featuring here
  • 16:49 - 16:52
    and we will look for questions
    and feedback from you
  • 16:52 - 16:58
    about what we think
    and what you think should be there
  • 16:58 - 17:02
    or how we should fix things
    if it doesn't work right.
  • 17:02 - 17:06
    Wikidata Game is a platform
    built by a community member Magnus,
  • 17:06 - 17:09
    a celebrity in this community, I think.
  • 17:10 - 17:13
    And by showing this
    we are providing datasets
  • 17:13 - 17:20
    but we also want to let people know
    that we are not reinventing the wheels,
  • 17:20 - 17:21
    that we are not trying to...
  • 17:21 - 17:24
    When we come up with some idea,
    we look into with community
  • 17:24 - 17:27
    and see if there is
    existing tools that's there
  • 17:27 - 17:30
    and how we can be
    a part of the ecosystem
  • 17:30 - 17:36
    rather than building everything
    independently and everything separately.
  • 17:37 - 17:39
    And this is the current status.
  • 17:40 - 17:43
    By early results, we show that Wikidata...
  • 17:45 - 17:47
    a few games that we released
  • 17:47 - 17:52
    have triggered and proved activity
    on the entities related
  • 17:53 - 17:55
    and a few follow up.
  • 17:55 - 17:57
    One thing that we have come up with,
  • 17:57 - 18:00
    as I have talked
    to a few community members
  • 18:00 - 18:02
    is the PreCheck idea
  • 18:02 - 18:09
    that is basically providing
    preliminary check about bulk uploads,
  • 18:09 - 18:12
    sampled preliminary check
    by community member
  • 18:12 - 18:14
    and use that to generate a report,
  • 18:14 - 18:16
    make it easier for discussions
  • 18:16 - 18:20
    about whether this big block
    of Wikidata datasets
  • 18:20 - 18:25
    should be included
    or uploaded to wikidata.org
  • 18:25 - 18:27
    or it should be rechecked or fixed.
  • 18:31 - 18:36
    And there is another project
    that is mostly a dataset project
  • 18:36 - 18:37
    called CatFacts.
  • 18:38 - 18:43
    CatFacts is datasets that we generate
  • 18:43 - 18:46
    about facts from categories,
  • 18:46 - 18:50
    the one that you see,
    the Christian Scientist, just now
  • 18:51 - 18:56
    is actually an interesting outlier
    of data points
  • 18:56 - 18:58
    from this effort.
  • 18:58 - 19:02
    This goal is to generate
    the facts from category
  • 19:02 - 19:07
    which we think have been
    very rich facts online that people...
  • 19:08 - 19:10
    that has been under leverage.
  • 19:10 - 19:14
    But before it can be fully leveraged
  • 19:14 - 19:17
    we need to make sure
    that quality is good enough as well
  • 19:17 - 19:22
    and there is efforts
    of putting it onto Wikidata Game
  • 19:22 - 19:24
    and there is effort that we're thinking
  • 19:24 - 19:27
    maybe building PreCheck
    would help as well.
  • 19:28 - 19:30
    And it's still in early stage.
  • 19:30 - 19:34
    Feel free to come to talk us
    about other efforts,
  • 19:34 - 19:38
    other ideas you think
    about datasets we could provide.
  • 19:38 - 19:42
    The Bot, which is communication tools.
  • 19:42 - 19:45
    We know that Bot can do many things
    like writhing Wikipedia article
  • 19:45 - 19:50
    but we promised
    that we don't write actual article
  • 19:50 - 19:53
    but we mostly use it
  • 19:53 - 19:58
    as a way to communicate
    from, let's say, user talk
  • 19:59 - 20:04
    to give us access
    to large scale conversations
  • 20:04 - 20:06
    with the community members.
  • 20:06 - 20:10
    Explorer is going to show
    all our datasets,
  • 20:10 - 20:12
    our toolings, their stats
  • 20:12 - 20:15
    and queries you can run on our things.
  • 20:15 - 20:18
    Stay tuned, this one is releasing soon.
  • 20:19 - 20:21
    And we have several other ideas
  • 20:21 - 20:24
    but I would jump
    to this overall portfolio.
  • 20:24 - 20:28
    It would be several projects
    to begin with datasets and tooling,
  • 20:28 - 20:30
    and what we are doing currently
  • 20:30 - 20:33
    is Explorer, Battlefield,
    CatFacts and PageRank,
  • 20:33 - 20:40
    and there are some other upcoming ideas
    like PreCheck, CitePool and Bubbles.
  • 20:41 - 20:46
    And this is one of the diagrams
  • 20:46 - 20:49
    that I want to show you.
  • 20:49 - 20:53
    We want to not only use
    one individual project
  • 20:53 - 20:55
    to contribute the community
  • 20:55 - 20:58
    and also generate the training data
    for the research, academia,
  • 20:58 - 21:01
    we also have an idea
  • 21:01 - 21:05
    that these projects may work together.
  • 21:06 - 21:09
    For example, the CitePool,
    the system that we want to build
  • 21:09 - 21:15
    to allow people to easier find citations
    for Wikipedia articles or Wikidata
  • 21:16 - 21:19
    but also use the Explorer
    to display the result--
  • 21:19 - 21:23
    it depends on the page rank
    scorances of datasets
  • 21:24 - 21:30
    to determine how to rank the citation page
    that we will recommend
  • 21:30 - 21:36
    and use the PreCheck
    to do quality, sanity check
  • 21:36 - 21:40
    and maybe create
    bulk batch reports by Bot
  • 21:40 - 21:44
    and the PreCheck will depend
    on the Game as well.
  • 21:51 - 21:53
    If some of our community friends
  • 21:53 - 21:55
    have been following
    the progress of WikiLoop,
  • 21:55 - 21:59
    we have been through ice-breaking phase,
  • 22:00 - 22:02
    we were trying to earn the community trust
  • 22:02 - 22:06
    because we know how cautious
    we need to be
  • 22:06 - 22:10
    coming to contribute to a movement
  • 22:10 - 22:15
    that relies so much
    on the neutrality and non-bias policies.
  • 22:15 - 22:20
    And we have gradually start to have ideas
  • 22:20 - 22:23
    about tools and datas
    and find the direction
  • 22:23 - 22:26
    of how we can possibly
    make this sustainable.
  • 22:26 - 22:32
    And we are looking into creating
    long-term sustainability,
  • 22:32 - 22:35
    both internally and also externally,
  • 22:35 - 22:39
    both in terms of getting resource
    and getting support,
  • 22:39 - 22:45
    also externally of getting engagement,
    getting usage, and getting contributors,
  • 22:46 - 22:48
    starting from next quarter.
  • 22:49 - 22:53
    I want to quote Evan You,
    who is a creator
  • 22:53 - 22:59
    of popular frontend framework Vue.js,
  • 22:59 - 23:01
    "Software development
    gets tremendously harder
  • 23:01 - 23:06
    when you start to have to convince people
    instead of just writing the code."
  • 23:06 - 23:09
    This applies to editing
    Wikipedia or Wikidata.
  • 23:09 - 23:13
    It's very easy to click a button
    and add individual articles
  • 23:13 - 23:19
    but also it's very hard
    when you need to convince people.
  • 23:23 - 23:27
    I hope to leave some time for questions,
  • 23:27 - 23:32
    although we only have few,
    probably one or two minutes.
  • 23:33 - 23:36
    Yes, so we have about two minutes.
  • 23:36 - 23:39
    So if people want to shout questions out,
    I'll bring the mic over.
  • 23:41 - 23:42
    Hands up maybe.
  • 23:45 - 23:50
    (person 1) So where would I go to
    at this moment if I would like to use this
  • 23:50 - 23:54
    to solve some of the problem
    with chemicals,
  • 23:54 - 23:57
    where some Wikipedia pages
    about chemicals,
  • 23:57 - 24:00
    they have a chem box
    about a specific chemical
  • 24:00 - 24:04
    but are otherwise about
    a class of chemicals.
  • 24:04 - 24:06
    Is that something
    where WikiLoop could help?
  • 24:08 - 24:13
    I think that's the individual
    domain expertize part, right?
  • 24:13 - 24:16
    If you are talking
    about topics of articles
  • 24:16 - 24:19
    that are associated with specific topics.
  • 24:19 - 24:21
    We are trying to...
    we might be able to help
  • 24:21 - 24:26
    but we are trying to tackle the problem
    that is like more general currently.
  • 24:26 - 24:33
    And overall the goal is
    to find the possibility of
  • 24:35 - 24:39
    empowering human beings productivity
  • 24:39 - 24:42
    and also trying to generate the knowledge
  • 24:42 - 24:44
    that potentially helps...
  • 24:44 - 24:47
    the training data that potentially
    helps the algorithms.
  • 24:50 - 24:52
    (person 2) I think we have time
    for a very quick one.
  • 24:55 - 24:59
    (person 3) Are you also going to do this
    for search of data on Commons?
  • 25:00 - 25:01
    Yeah, we hope to...
  • 25:01 - 25:05
    If you are referring to Battlefield
    or counter-vandalism tools,
  • 25:06 - 25:12
    yeah, we are planning
    to expand it to other Wiki projects,
  • 25:12 - 25:14
    including Commons in Wikidata.
  • 25:15 - 25:17
    (person 2) I think that's all the questions
    we have time for
  • 25:17 - 25:20
    but if you'd like to show
    your appreciation for [Victor.]
  • 25:20 - 25:21
    Thank you.
  • 25:21 - 25:25
    (applause)
Title:
cdn.media.ccc.de/.../wikidatacon2019-1147-eng-Project_WikiLoop_hd.mp4
Video Language:
English
Duration:
25:30

English subtitles

Revisions Compare revisions