Table 2:
Important notations and denotations.
NotationDenotation
$w→$ Vector corresponding to a word w
$wd→$ Debiased version of $w→$
V Vocabulary set
Vp The set of words which are preserved during the debiasing procedure
Vd The set of words which are subjected to the debiasing procedure
D Set of dictionaries
di A particular dictionary from the set D
$g→$ Gender direction
$Db(w→)$ Direct bias of a word w
$β(w1→,w2→)$ Indirect bias between a pair of words w1 and w2
$η(w→)$ Gender-based proximity bias of a word w
Nw Set of neighboring words of a word w
$Fr(wd→)$ Repulsion objective function
$Fa(wd→)$ Attraction objective function
$Fn(wd→)$ Neutralization objective function
$F(wd→)$ Multi-objective optimization function
KBC Knowledge Based Classifier
BBN Bias Based Network
GIPE Gender-based Illicit Proximity Estimate
NotationDenotation
$w→$ Vector corresponding to a word w
$wd→$ Debiased version of $w→$
V Vocabulary set
Vp The set of words which are preserved during the debiasing procedure
Vd The set of words which are subjected to the debiasing procedure
D Set of dictionaries
di A particular dictionary from the set D
$g→$ Gender direction
$Db(w→)$ Direct bias of a word w
$β(w1→,w2→)$ Indirect bias between a pair of words w1 and w2
$η(w→)$ Gender-based proximity bias of a word w
Nw Set of neighboring words of a word w
$Fr(wd→)$ Repulsion objective function
$Fa(wd→)$ Attraction objective function
$Fn(wd→)$ Neutralization objective function
$F(wd→)$ Multi-objective optimization function
KBC Knowledge Based Classifier
BBN Bias Based Network
GIPE Gender-based Illicit Proximity Estimate
Close Modal