Analyze, Detect and Remove Gender Stereotyping from Bollywood Movies
VI) Bias detection using Actions - At this
point we have the actions extracted from biased
data corresponding to each gender. We can now
use this data against fact data to check for bias.
We will describe in the following system walk-
through section how we use it on-the-fly to check
for bias.
VII) Bias Removal - We construct a knowl-
edge graph for each cast using relations from
Stanford dependency parser. We use this graph
to calculate the between-ness centrality for each
cast and store these centrality scores in a knowl-
edge base. We use the between-ness centrality
score to interchange genders after we detect the
bias.
5. Walk-through using an example
The system DeCogTeller takes in a text input
from the user. The user starts entering a bi-
ased movie plot text for a movie, say, “Kaho na
Pyar Hai” in Figure 14. This natural language
text is submitted into the system in which, first,
the text is co-referenced using OpenIE. Then, us-
ing IBM NLU API and UIUC Semantic Role
Labeller actions pertaining to each cast are ex-
tracted and these are checked with gender specific
and gender neutral lists. If for a corresponding
cast gender,action pair the corresponding vector
is located in gender specific list then it can not
be termed as a biased action. But on the other
hand if a cast gender,action pair occurring in the
plot is not found in gender-specific but the oppo-
site gender is found in gender-neutral list, then
we tag the statement as a biased statement.
As an example text if the user enters - “Ro-
hit is an aspiring singer who works as a salesman
in a car showroom, run by Malik (Dalip Tahil).
One day he meets Sonia Saxena (Ameesha Patel),
daughter of Mr. Saxena (Anupam Kher), when
he goes to deliver a car to her home as her birth-
day present.” At the very fist step, co-referencing
is done which coverts the above text to - “Rohit
is an aspiring singer who works as a salesman in a
car showroom, run by Malik (Dalip Tahil). One
day Rohit meets Sonia Saxena (Ameesha Patel),
daughter of Mr. Saxena (Anupam Kher), when
Rohit goes to deliver a car to her home as her
birthday present.” After this step, we extract ac-
tions corresponding to each cast and then check
for bias. Here corresponding to cast Rohit we
have the following actions - {singer, salesman,
meets, deliver}. The gender for Rohit is detected
by using wiki page of Hritik Roshan and is la-
belled as “male”. We find actions correspond-
ing to cast Sonia and find the following actions-
{daughter-of}. Then we run our gender-specific
and gender neutral checks and find that the ac-
tions are gender neutral. Hence there is a bias
that exists. We do the similar thing for other
cast members. Then, at the background, we ex-
tract highest centrality male and highest central-
ity female. And then switch their gender to gen-
erate de-biased plot. Figure 15 shows the de-
biased plot. Also, there is an option given to the
user to view the knowledge graphs for biased text
and unbiased text to see how nodes in knowledge
graph change.
6. Discussion and Ongoing Work
While our analysis points towards the presence of
gender bias in Hindi movies, it is gratifying to see
that the same analysis was able to discover the
slow but steady change in gender stereotypes.
We would also like to point out that the goal
of this study is not to criticize one particular do-
main. Gender bias is pervasive in all walks of life
including but not limited to the Entertainment
Industry, Technology Companies, Manufacturing
Factories & Academia. In many cases, the bias
is so deep rooted that it has become the norm.
We truly believe that the majority of people dis-
playing gender bias do it unconsciously. We hope
that ours and more such studies will help people
realize when such biases start to influence every
day activities, communications & writings in an
unconscious manner, and take corrective actions
to rectify the same. Towards that goal, we are
building a system which can re-write stories in
a gender neutral fashion. To start with we are
focusing on two tasks:
a) Removing Occupation Hierarchy : It
is common in movies, novel & pictorial depiction
to show man as boss, doctor, pilot and women
as secretary, nurse and stewardess. In this work,
we presented occupation detection. We are ex-
tending this to understand hierarchy and then
evaluate if changing genders makes sense or not.
For example, while interchanging ({male, doc-
tor}, {female, nurse}) to ({male, nurse}, {female,
doctor}) makes sense but interchanging {male,
11