Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Antisocial Computing

Antisocial Computing

In online social networks, large information cascades can develop as people share content with one another. However, as these cascades develop through complex processes, prior work has argued that their future trajectory may be inherently unpredictable. My research introduces methods for studying the mechanisms of these cascades and predicting their spread. Analyzing billions of interactions by hundreds of millions of users on Facebook, I show how the future growth and structure of these cascades can be predicted, how cascades may resurface after lying dormant for months, and how diverse social protocols can produce large information cascades. Through revealing the mechanisms in which information diffuses in social media, this work explores a future where systems can better promote sharing behavior online.

Presented at the Stanford Human-Computer Interaction Seminar.

Justin Cheng

June 02, 2017
Tweet

More Decks by Justin Cheng

Other Decks in Research

Transcript

  1. Justin Cheng
    Stanford University
    ANTISOCIAL COMPUTING
    Explaining and Predicting Negative Behavior Online

    View full-size slide

  2. Vieweg, et al. (2010); Kittur, et al. (2013); Burke & Kraut (2016)

    View full-size slide

  3. Time (2016); The Atlantic (2016); Vanity Fair (2017)

    View full-size slide

  4. 47% of online users
    have been harassed
    Data & Society (2017)

    View full-size slide

  5. Popular Science (2013); The Verge (2015); Chicago Sun-Times (2014)

    View full-size slide

  6. Why is bad behavior so prevalent?

    (›°□°)›ớ ᵲᴸᵲ
    Research Question

    View full-size slide

  7. Understanding bad behavior helps
    us build healthier communities
    Implications
    Systems
    Guidelines Interventions

    View full-size slide

  8. Antisocial behavior is largely
    due to sociopaths
    Prior Work
    Donath (1999); Hardaker (2010); Buckels, et al. (2014)

    View full-size slide

  9. Antisocial behavior is largely
    due to ordinary people
    This Work

    View full-size slide

  10. Antisocial Behavior & Its Spread
    Talk Outline
    What causes antisocial behavior?
    Can such cascades be predicted?
    Does it worsen over time?
    1
    2
    3

    View full-size slide

  11. Data Mining + Crowdsourcing
    Research Approach
    Large-scale Analysis + Experiments

    View full-size slide

  12. Identifying principles of
    online behavior
    The Broader Picture
    Data + ML + Network Science + HCI

    View full-size slide

  13. Antisocial Behavior & Its Spread
    Talk Outline
    1
    2
    3
    What causes antisocial behavior?
    Can such cascades be predicted?
    Does it worsen over time?

    View full-size slide

  14. CSCW 2017 (Best Paper); ICWSM 2015 (Honorable Mention)
    with M. Bernstein, C. Danescu-Niculescu-Mizil, J. Leskovec
    CAN ANYONE
    BECOME A TROLL?
    Causes of Antisocial Behavior in Online Discussions

    View full-size slide

  15. CONTENT WARNING!

    This talk contains depictions of trolling that use strong language.
    !

    View full-size slide

  16. It also shows that Islam and
    Christianity teaching women to
    dress modest could be right
    afterall.

    View full-size slide

  17. It also shows that Islam and
    Christianity teaching women to
    dress modest could be right
    afterall.
    Clearly that is the only logical
    conclusion to this article. Now if
    you'll excuse me, I need to iron
    my tarp. I have work on Monday,
    and I want to appear 'modest'.
    fail at life. go bomb yourself.
    Religious nut alert

    View full-size slide

  18. We studied multiple large comment-
    based news communities.
    470M posts 831M votes
    76M users

    View full-size slide

  19. What is trolling?

    View full-size slide

  20. What is trolling?
    Engaging in negatively marked online
    behavior
    Taking pleasure in upsetting others
    Not following the rules
    Disrupting a group while staying
    undercover
    Donath (1999); Hardaker (2010); Kirman (2012); Schwartz (2008)

    View full-size slide

  21. Trolling is behavior that occurs
    outside community norms.
    Defined using community guidelines
    Our Definition
    e.g., name-calling, personal attacks, profanity, threats, hate speech, ethnically/racially offensive material

    View full-size slide

  22. Are trolls just a vocal minority?
    Donath (1999); Hardaker (2010); Shachaf & Hara (2010); NYT (2008); Wired (2014); Vox (2014)

    View full-size slide

  23. How much do trolls troll?
    Proportion of Banned Users
    0
    0.1
    0.2
    0.3
    0.4
    Proportion of Deleted Posts
    0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

    View full-size slide

  24. The distribution of trolls is bimodal
    Proportion of Banned Users
    0
    0.1
    0.2
    0.3
    0.4
    Proportion of Deleted Posts
    0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

    View full-size slide

  25. Are there two types of trolls?
    Proportion of Banned Users
    0
    0.1
    0.2
    0.3
    0.4
    Proportion of Deleted Posts
    0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
    Situational trolling?
    Lifelong trolling?

    View full-size slide

  26. What if antisocial behavior is
    situational?

    View full-size slide

  27. Challenge: how to show that
    antisocial behavior is situational?
    Observational data isn’t causal

    View full-size slide

  28. Challenge: how to show that
    antisocial behavior is situational?
    Experiments hard to generalize

    View full-size slide

  29. Simulated Discussion Experiment Large-Scale Analysis
    Solution: Experiment + Observational Study

    View full-size slide

  30. Anyone can become a troll
    Our Hypothesis

    View full-size slide

  31. “Broken windows” theory
    Zimbardo (1969); Wilson (1982)

    View full-size slide

  32. Unpleasant stimuli increase aggression
    Jones & Bogat (1978); Rotton & Frey (1985)

    View full-size slide

  33. Experiment: simulated discussion forum

    View full-size slide

  34. N=667, 40% female
    Quiz Discussion
    Experimental method

    View full-size slide

  35. Quiz Discussion
    Experimental method
    ×

    View full-size slide

  36. Positive/Negative Mood Positive/Negative Context
    Experimental method
    ×
    Quiz Discussion

    View full-size slide

  37. Easy quiz (positive mood)

    View full-size slide

  38. Difficult quiz (negative mood)

    View full-size slide

  39. Positive discussion context

    View full-size slide

  40. Negative discussion context

    View full-size slide

  41. People reported being in a worse
    mood after the difficult quiz
    Easy: 12.2 Difficult: 40.8
    (POMS mood disturbance [higher scores = worse mood], p<0.01)
    Manipulation Check

    View full-size slide

  42. Initial seed posts in the negative
    context condition perceived worse
    Positive: 90% upvoted Negative: 36% upvoted
    (p < 0.01)
    Manipulation Check

    View full-size slide

  43. How did trolling differ across
    conditions?
    Two expert raters labeled posts independently

    View full-size slide

  44. How did trolling differ across conditions?
    Positive
    Mood
    Negative
    Mood
    Positive
    Context
    Negative
    Context
    % Troll Posts

    View full-size slide

  45. Trolling is lowest with positive conditions…
    Positive
    Mood
    Negative
    Mood
    Positive
    Context
    35%
    Negative
    Context
    % Troll Posts

    View full-size slide

  46. …increases with either negative condition…
    Positive
    Mood
    Negative
    Mood
    Positive
    Context
    35% 49%
    Negative
    Context
    47%
    % Troll Posts

    View full-size slide

  47. …and almost doubles in the worst case
    Positive
    Mood
    Negative
    Mood
    Positive
    Context
    35% 49%
    Negative
    Context
    47% 68%
    % Troll Posts
    (p < 0.05 using a mixed effects logistic regression model)

    View full-size slide

  48. Negative affect almost triples
    Positive
    Mood
    Negative
    Mood
    Positive
    Context
    1.1% 1.4%
    Negative
    Context
    2.3% 2.9%
    % Negative Affect Words (LIWC)
    (p < 0.05)

    View full-size slide

  49. Hilary is a solid candidate. As a woman, I
    appreciate that she's a woman, but it's not
    the only reason I think she would do well in
    office.
    Positive Mood + Context

    View full-size slide

  50. Anyone who votes for her is a complete
    idiot. These supporters are why this country
    is in such bad shape now. Uneducated
    people.
    Negative Mood + Context

    View full-size slide

  51. Bad mood and negative discussion
    context increase trolling

    View full-size slide

  52. Simulated Discussion Experiment Large-Scale Analysis of CNN.com
    Online Experiment + Observational Study

    View full-size slide

  53. Can trolling, like mood, vary with
    the time of day and day of week?
    Replicating Mood
    Golder & Macy (2011)
    Neg. Affect
    Time of Day
    ?

    View full-size slide

  54. How does trolling vary with time of day?
    Proportion of Flagged Posts
    0.03
    0.033
    0.036
    0.039
    0.042
    Time of Day
    0 6 12 18 24

    View full-size slide

  55. Trolling peaks in the evening…
    Proportion of Flagged Posts
    0.03
    0.033
    0.036
    0.039
    0.042
    Time of Day
    0 6 12 18 24
    Negative Affect
    (Golder & Macy)

    View full-size slide

  56. …and early in the work week.
    Proportion of Flagged Posts
    0.03
    0.032
    0.034
    0.036
    0.038
    0.04
    Day of Week
    Mon Tue Wed Thu Fri Sat Sun

    View full-size slide

  57. Trolling peaks when moods are worse
    Time of Day
    Proportion of
    Flagged Posts
    Negative Affect
    Proportion of
    Downvotes
    Day of Week

    View full-size slide

  58. Mood spills over from prior
    discussions
    Replicating Mood

    View full-size slide



  59. (Discussion)
    Mood spills over from prior
    discussions

    View full-size slide




  60. ?

    (Discussion)
    (Unrelated Discussions)

    ?

    Mood spills over from prior
    discussions

    View full-size slide





  61. (Discussion)


    Mood spills over from prior
    discussions

    View full-size slide

  62. A user who trolled in a previous
    discussion is twice as likely to troll
    in a later, unrelated discussion
    (p < 0.01)
    Replicating Mood

    View full-size slide

  63. The initial post affects subsequent
    trolling
    Replicating Context

    View full-size slide

  64. (Separate discussions of same article)

    ? ? ?

    ? ? ?
    The initial post affects subsequent
    trolling

    View full-size slide

  65. (Separate discussions of same article)


    The initial post affects subsequent
    trolling

    View full-size slide

  66. An initial troll post increases the
    subsequent trolling by 63%
    (p < 0.01)
    Replicating Context

    View full-size slide

  67. Can we predict trolling
    before it happens?
    Balanced dataset of 120K posts Logistic regression

    View full-size slide

  68. Mood Context The User
    What factors affect trolling?
    Trolling is situational Trolling is innate

    View full-size slide

  69. How predictable are troll posts?
    User-specific
    Mood
    Discussion Context
    Combined
    AUC
    0.00 0.20 0.40 0.60 0.80
    0.78
    0.74
    0.60
    0.66

    View full-size slide

  70. How predictable are troll posts?
    User-specific
    Mood
    Discussion Context
    Combined
    AUC
    0.00 0.20 0.40 0.60 0.80
    0.78
    0.74
    0.60
    0.66

    View full-size slide

  71. How predictable are troll posts?
    User-specific
    Mood
    Discussion Context
    Combined
    AUC
    0.00 0.20 0.40 0.60 0.80
    0.78
    0.74
    0.60
    0.66

    View full-size slide

  72. Troll or not?
    User

    View full-size slide

  73. Troll or not?
    User
    Mood

    View full-size slide

  74. Troll or not?
    User
    Mood Other users
    { }

    View full-size slide

  75. Because trolling is situational,
    ordinary people can end up trolling

    View full-size slide

  76. Can voting mitigate bad behavior?

    View full-size slide

  77. Downvoting causes
    negative behavior to worsen
    Our Hypothesis

    View full-size slide

  78. Antisocial Behavior & Its Spread
    Talk Outline
    1
    2
    3
    What causes antisocial behavior?
    Can such cascades be predicted?
    Does it worsen over time?

    View full-size slide

  79. How Antisocial Behavior Worsens
    ICWSM 2014
    with C. Danescu-Niculescu-Mizil, J. Leskovec
    CAN ANTISOCIAL
    BEHAVIOR SPI
    R
    A
    L?

    View full-size slide

  80. Downvoting causes
    negative behavior to worsen
    Our Hypothesis

    View full-size slide

  81. What effects do evaluations have?
    Positively
    evaluated
    ?
    ?
    Negatively
    evaluated

    View full-size slide

  82. What is a positive or negative evaluation?

    View full-size slide

  83. Defining positive and negative evaluations
    :
    9 1
    2 8
    N↑

    N↑
    + N↓
    2

    2+8
    = = 0.2 p↓

    : N↑

    N↑
    + N↓
    9

    9+1
    = = 0.9 p↑

    Positive Evaluation
    Negative Evaluation
    (validated using a crowdsourcing experiment)

    View full-size slide

  84. What effects do evaluations have?
    Positively
    evaluated
    ?
    ?
    Negatively
    evaluated

    View full-size slide

  85. Does feedback encourage better behavior?
    Skinner (1938)

    View full-size slide

  86. Or is bad stronger than good?
    Brinko (1993); Baumeister, et al. (2001)

    View full-size slide

  87. Four large comment-based news
    communities

    View full-size slide

  88. What effects do evaluations have?
    Positively
    evaluated
    Negatively
    evaluated

    View full-size slide

  89. What effects do evaluations have?




    View full-size slide

  90. What effects do evaluations have?




    Before After
    vs.
    Before After
    vs.

    View full-size slide

  91. Challenge: how to compare
    different users and posts?
    Aren’t downvoted users/posts inherently worse?

    View full-size slide

  92. Solution: propensity score matching
    PSM: Rosenbaum (1983); CEM: Iacus, et al. (2012)
    Positively
    evaluated
    Negatively
    evaluated

    View full-size slide

  93. Match on text quality
    Similar text
    quality q
    }

    View full-size slide

  94. Computing text quality
    Learn p with bigrams

    (binomial regression)
    1 3
    lorem ipsum…
    q = ?
    Lorem…
    ? ?
    9 2
    lorem ipsum…

    Text quality q is
    predicted p

    View full-size slide

  95. Validating text quality
    Manually label subset (n=171) using crowdsourcing
    lorem ipsum…
    Good
    Bad
    Good
    Good
    # Good

    # Total
    q’=

    View full-size slide

  96. Validating text quality
    Manually label subset (n=171) using crowdsourcing
    lorem ipsum…
    Good
    Bad
    Good
    Good
    # Good

    # Total
    q’=
    ? ?
    corr(q’, p) = corr(q’, q) =

    View full-size slide

  97. Validating text quality
    (n.s.) (p < 0.01)
    p: actual proportion of upvotes
    q: text quality (predicted proportion)
    q’: crowd guess
    corr(q’, q) =
    corr(q’, p) = 0.11 0.25

    View full-size slide

  98. Validating text quality
    |Residuals|
    0
    0.2
    0.4
    0.6
    0.8
    1
    0.6 0.7 0.8 0.9 1
    |q − q’|
    (n.s. using a Breusch-Pagan test)
    Text Quality q

    View full-size slide

  99. Match on text quality
    Similar text
    quality
    q(c↑
    )=q(c↓
    )
    }

    View full-size slide

  100. …as well as other covariates
    Similar history
    (# posts, overall
    proportion of
    upvotes, etc.)
    { ≈



    View full-size slide

  101. …as well as other covariates






    View full-size slide

  102. How are subsequent posts evaluated?






    View full-size slide

  103. How much are evaluations due to
    textual or community effects?

    View full-size slide

  104. How much are evaluations due to
    textual effects (i.e., people writing worse)?
    f***ing a******
    i.e., downvoting because of post content

    View full-size slide

  105. How much are evaluations due to
    community effects (i.e., inherent bias)?
    We dislike you.
    i.e., downvoting because of community dislikes author
    We dislike you.

    View full-size slide

  106. Do people write better/worse after
    a positive/negative evaluation?
    Textual Effects

    View full-size slide






  107. Better/Worse?
    Do people write better/worse after
    a positive/negative evaluation?

    View full-size slide

  108. Text quality drops significantly after
    a negative evaluation…
    (p < 0.05, mean effect size r = 0.18)
    … …
    Negativity bias

    View full-size slide

  109. …but doesn’t change after
    a positive evaluation
    … …
    (n.s.)
    Negativity bias

    View full-size slide

  110. How does community bias
    change after an evaluation?
    Community Effects

    View full-size slide

  111. … …
    … …
    How does community bias
    change after an evaluation?

    View full-size slide

  112. Measuring community bias
    N↑ q
    N↓
    N↑

    N↑
    + N↓
    = 0.5
    p(c) =
    = 0.8
    q(c)
    p(c) q(c) = 0.3
    Prop. Upvotes
    Text Quality
    Community Bias − −

    View full-size slide

  113. Community bias increase more after
    a negative than positive evaluation
    (p < 0.01, mean effect size r = 0.13)
    … …
    Halo effect

    View full-size slide

  114. More positive
    Negative Eval.
    Positive Eval.

    View full-size slide

  115. More positive
    Negative Eval.
    Positive Eval.
    Similar text quality

    q(c↑
    )=q(c↓
    )

    View full-size slide

  116. More positive
    Before Negative Eval.
    Positive Eval.
    Similar history Similar text quality

    q(c↑
    )=q(c↓
    )

    View full-size slide

  117. More positive
    Before After
    Negative Eval.
    Positive Eval.
    Similar history Similar text quality

    q(c↑
    )=q(c↓
    )
    Worse text quality

    q(c↑(1,3)
    ) > q(c↓(1,3)
    )
    *

    View full-size slide

  118. More positive
    Before After
    Negative Eval.
    Positive Eval.
    Similar history Similar text quality

    q(c↑
    )=q(c↓
    )
    Worse text quality

    q(c↑(1,3)
    ) > q(c↓(1,3)
    )
    Worse perception

    q(c↓(1,3)
    ) - p(c↓(1,3)
    ) >
    q(c↑(1,3)
    ) - p(c↑(1,3)
    )
    *
    *

    View full-size slide

  119. What happens to
    negatively-evaluated users?

    View full-size slide

  120. They post worse content
    Perceptions of them become worse
    They post more frequently*
    They evaluate others more negatively*
    * More details in our ICWSM 2014 paper (http://bit.ly/feedback-paper)
    What happens to
    negatively-evaluated users?

    View full-size slide

  121. Troll or not?
    User
    Mood Other users
    { }

    View full-size slide

  122. Troll or not?
    User
    Mood Other users
    { }

    View full-size slide

  123. Troll or not?
    User
    Mood Other users
    { }

    View full-size slide

  124. Trolls may start out normal, but tip
    into a spiral and never recover

    View full-size slide

  125. Do communities worsen over time?

    View full-size slide

  126. Communities may worsen over time (?)
    Proportion of Upvotes
    0.6
    0.65
    0.7
    0.75
    0.8
    Time
    December 2012 February 2013 April 2013 June 2013 August 2013

    View full-size slide

  127. Antisocial Behavior & Its Spread
    Talk Outline
    1
    2
    3
    What causes antisocial behavior?
    Can such cascades be predicted?
    Does it worsen over time?

    View full-size slide

  128. The Predictability of Information Cascades in Social Networks
    CAN C S
    BE PREDICTED?
    AS A
    C DE
    WWW 2014; WWW 2016
    with L. Adamic, P. A. Dow, J. Kleinberg, J. Leskovec

    View full-size slide

  129. Rumors on Facebook
    ICWSM 2014 (with A. Friggeri, L. Adamic, and D. Eckles)

    View full-size slide

  130. Same rumor, different popularity

    View full-size slide

  131. Are these cascades predictable?

    View full-size slide

  132. Are cascades unpredictable?

    View full-size slide

  133. Large cascades are rare
    Empirical CCDF
    0.00
    0.20
    0.40
    0.60
    0.80
    1.00
    Cascade size
    0 200 400 600 800 1000
    0.09
    100

    View full-size slide

  134. “Increasing the strength of social
    influence increased both inequality
    and unpredictability of success.”
    Salganik, Dodds & Watts (2006)

    View full-size slide

  135. Cascades can recur after long periods

    View full-size slide

  136. Cascades are predictable
    Our Hypothesis
    size, structure, content even if they recur

    View full-size slide

  137. How do we begin to predict
    cascade growth?

    View full-size slide

  138. Challenge: how to predict cascade growth?
    ?
    k=5 reshares

    observed

    View full-size slide

  139. Challenge: how to predict cascade growth?
    ?
    Will a cascade get 100 reshares?
    Exactly how big will a small cascade get?
    Only consider the largest cascades?

    View full-size slide

  140. Challenge: how to predict cascade growth?
    ?
    Will a cascade get 100 reshares?
    Exactly how big will a small cascade get?
    Only consider the largest cascades?
    class imbalance
    outliers skew results
    selection bias

    View full-size slide

  141. Solution: will a cascade reach the median?
    ? ≤ the median f(k)
    ≥ the median f(k)
    k=5 reshares

    observed

    View full-size slide

  142. Solution: will a cascade double in size?
    ? ≤ the median f(k)
    ≥ the median f(k)
    k=5 reshares

    observed

    View full-size slide

  143. Given that a cascade has obtained
    k reshares, will it double in size?
    balanced track growth over time
    Cascade Growth Prediction Problem

    View full-size slide

  144. Reshare cascades on Facebook
    70M cascades 5B reshares Activity over 28 days

    View full-size slide

  145. Content

    has overlaid text
    captions

    User

    friend count

    gender

    Structural

    tree depth
    outdegree

    Temporal

    time between shares
    change in time

    What factors affect predictability?

    View full-size slide

  146. How predictable is cascade doubling?
    All
    Temporal
    All but temporal
    Structural
    User
    Content
    AUC (k=5)
    0.00 0.23 0.45 0.68 0.90
    0.58
    0.71
    0.74
    0.79
    0.87
    0.88
    All but temporal

    View full-size slide

  147. Given that a cascade has obtained
    k reshares, will it double in size?
    Cascade Growth Prediction Problem

    View full-size slide

  148. Given that a cascade has obtained
    5 reshares, will it double in size?
    Cascade Growth Prediction Problem

    View full-size slide

  149. Given that a cascade has obtained
    100 reshares, will it double in size?
    Cascade Growth Prediction Problem

    View full-size slide

  150. How does performance change with k?
    k = 5 k > 10
    k = 10 k > 20

    View full-size slide

  151. How does performance change with k?
    k = 5 k > 10
    k = 10 k > 20
    Less data
    More data

    View full-size slide

  152. How does performance change with k?
    k = 5 k > 10
    k = 10 k > 20
    Shorter-term
    Longer-term

    View full-size slide

  153. Easier to predict larger cascades doubling
    Accuracy
    0.78
    0.79
    0.8
    0.81
    0.82
    Number of reshares observed, k
    0 25 50 75 100

    View full-size slide

  154. Cascade growth is predictable

    View full-size slide

  155. Cascade structure is predictable
    AUC = 0.80 for predicting structural virality
    * More details in our WWW 2014 paper (http://bit.ly/memes-paper)
    vs.

    View full-size slide

  156. Cascade recurrence is predictable
    AUC = 0.89 for predicting a subsequent burst
    * More details in our WWW 2016 paper (http://bit.ly/cascades-paper)
    vs.

    View full-size slide

  157. Troll or not?
    User
    Mood Other users
    { }

    View full-size slide

  158. What we now know
    What we thought
    Trolls are a
    vocal minority
    Trolls can be
    ordinary people
    Trolling is innate
    Trolling can spiral from
    a single bad post
    Cascades can be
    predicted
    Cascades are
    unpredictable
    ANTISOCIAL COMPUTING

    View full-size slide

  159. Predicting the demise of communities
    Proportion of Upvotes
    0.6
    0.65
    0.7
    0.75
    0.8
    Time
    December February April June August
    Future Directions

    View full-size slide

  160. Designing prosocial discussion platforms
    Future Directions

    View full-size slide

  161. Designing prosocial discussion platforms
    Future Directions

    View full-size slide

  162. Munger (2016)
    Introducing conversation mediators
    Future Directions
    Don’t be a n****r.
    Hey man, just remember that there
    are real people who are hurt
    when you harass them with that
    kind of language.
    (e.g., bots)

    View full-size slide

  163. WWW 2017 (with S. Kumar, J. Leskovec, and V.S. Subrahmanian)
    Identifying different types of trolling
    Future Directions
    Possibly the best blog I’ve ever
    read major props to you
    Thanks. I knew Marvel fans would try
    to flame me, but they have nothing
    other than “oh that’s your opinion”
    Quit talking to yourself […]
    (e.g., sockpuppets)

    View full-size slide

  164. Addressing polarization
    Future Directions
    Measuring algorithmic impact
    Tracking cascades at scale

    View full-size slide

  165. Holistic approaches for analyzing
    and building social systems
    Research Approach

    View full-size slide

  166. Holistic approaches for analyzing
    and building social systems
    Large-scale Analysis Experimentation
    +
    Macro-scale Micro-scale
    +
    Understand Build
    +
    Research Approach

    View full-size slide

  167. Multi-methods analyses identify patterns
    in data, verify hypotheses, make
    predictions, and develop social systems.
    Multi-methods analyses identify
    patterns in data, verify hypotheses,
    make predictions, and inform the
    design of better social systems.

    View full-size slide

  168. Jure Leskovec Michael Bernstein Jon Kleinberg Lada Adamic
    Thank you!

    View full-size slide

  169. James Landay Jeff Hancock
    Cristian
    Danescu-Niculescu-Mizil
    Dan Cosley
    Thank you!

    View full-size slide

  170. Thank you!
    Stanford HCI Group
    SNAP Group
    Stanford VPGE
    Microsoft Research
    Facebook
    Pinterest
    Disqus

    View full-size slide

  171. Justin Cheng / @jcccf / clr3.com
    Stanford University
    More resources and credits: http://bit.ly/jobtalkcredits
    ANTISOCIAL COMPUTING
    Explaining and Predicting Negative Behavior Online

    View full-size slide