12. Blind application of machine
learning algorithms may amplify
biases present in the data
BOLUKBASI ET AL. (2016)
Man is to computer programmer as woman is to homemaker?
debiasing word embeddings
30. Accountability
“The question of whether Machines Can
Think... is about as relevant as the question of
whether Submarines Can Swim.” ―Edsger Dijkstra
40. .Amy
. Joan
. Lisa
. Sarah
. Diana
. Kate
. Ann
. Donna
IAT (attributes list)
.John
. Paul
. Mike
. Kevin
. Steve
. Greg
. Jeff
. Bill
Attribute list 1 Attribute list 2
52. Word Embedding (evaluation)
Man is to King, as woman is to ____
King(vec) - Man(vec) + Woman(vec) = Queen(vec)
Rio is to Brazil, as Paris is to ____
Rio(vec) - Brazil(vec) + Paris(vec) = France(vec)
53. Man is to King, as woman is to ____
King(vec) - Man(vec) + Woman(vec) = Queen(vec)
Man is to King, as woman is to ____
King(vec) - Man(vec) + Woman(vec) = Lioness(vec)
Word Embedding (evaluation)
54. Man is to King, as woman is to ____
King(vec) - Man(vec) + Woman(vec) = Queen(vec)
Man is to doctor, as woman is to ____
doctor(vec) - Man(vec) + Woman(vec) = Nurse(vec)
Word Embedding (bias)
63. Cultural biases are expressed in
people’s language
CALISKAN ET AL. (2017)
Semantics derived automatically from language corpora contain human-like biases
67. Cohen's D Interpretation
0.01 very small
0.20 small
0.50 medium
0.80 large
1.20 very large
2.00 huge
Cohen’s D interpretation
68. Null Hypotesis testing
In statistical hypothesis testing, the alternative
hypothesis and the null hypothesis are the two rival
hypotheses which are compared by a statistical hypothesis
test.
72. CALISKAN ET AL. (2017)
Semantics derived automatically from language corpora contain human-like biases
Hypotesis testing
73. CALISKAN ET AL. (2017)
Semantics derived automatically from language corpora contain human-like biases
Hypotesis testing XKCD (201?)
https://www.xkcd.com/1478/
74. CALISKAN ET AL. (2017)
Semantics derived automatically from language corpora contain human-like biases
P-Value / Cohen’s d
CALISKAN ET AL. (2017)
Semantics derived automatically from language corpora contain human-like biases
Dist1 Dist2
77. Hate Speech
Hate speech can then be defined as the vilification of a
group’s Identity in order to oppress its members and deny
them equal rights.
Cherian ET AL. (2016)
Hate Spin: The Manufacture of Religious Offense and Its Threat to Democracy
78. Group Identity
Hate speech communicates extremely negative ideas about a
group, or a representative of that group, as defined by
identity markers such as race, religion and sexual
orientation.
Cherian ET AL. (2016)
Hate Spin: The Manufacture of Religious Offense and Its Threat to Democracy
84. Compare Languages with BEAT
1.02
1.73
…
Trump
Sexuality bias
Religion bias
Cohen’s D
0.35
0.62
…
English
Cohen’s D
85. Explore relations in vector space
Trump is to White Supremacists, as
Hillary is to ____
White_Supremacists(vec) - Trump(vec) + Hillary(vec) = ?(vec)
86. 0.76
0.1
…
Real Word values into online world?
.Modern Liberalism
.Gender equality
.Freedom of religion
…
Democratic Party’s
BEAT
Sexuality bias
Religion bias
87. Algorithm Bias
Fairness in natural language processing
Raphael Ottoni
<raphael@hekima.com.br>, <rapha@dcc.ufmg.br>
88. IAT e WEAT usam cohen’d
A premissa para que o mapeamento IAT -> WEAT é que o
tempo de resposta é equivalente a similaridade de cosseno
no embedding
É Preciso provar que o WEAT pode ser usado em diferentes
contexto, além dos testados no paper origianl
89. Algorithm Bias
Fairness in natural language processing
Raphael Ottoni
<raphael@hekima.com.br>, <rapha@dcc.ufmg.br>
90. Of Language and Values
Characterizing political groups by its ideology biases