By the age of 6,girls are less
likely than boys to label people
of their own gender as “really,
Paper reference: Gender stereotypes about
intellectual ability emerge early and influence
children’s interests (Science Journal)
Paper reference: “Man Is to
Computer Programmer as
Woman is to Homemaker?
Debiasing Word Embeddings”
Word2vec,a popular framework to
represent text data as vectors:
• man:king :: woman:queen
• sister:woman :: brother:man
• man:boss :: woman:receptionist
• man:programmer :: woman:homemaker
Paper reference: “Men Also Like
Shopping: Reducing Gender Bias
Amplification using Corpus-level
• The visual semantic labeling (vSRL)
tasks require large quantities of
labeled data,predominantly retrieved
from the web.
• 67% of cooking images in datasets have
a woman in the agent role.
• Machine-learning software trained on
the datasets didn’t just mirror those
biases,it amplified them: they do not
recognize images of men cooking.
Work promotion often
starts in casual contexts
Paper reference: “Automated
Experiments on Ad Privacy
Settings. A Tale of Opacity,
After creating 1,000 fictitious
users (50% women,50% men):
men saw 1,800 times a Google's
ad of a high-paying job. The
women saw it 300 times.
Facebook is letting Job advertisers target only men
A review by ProPublica found that 15 employers in the past
year,including Uber,have advertised jobs on Facebook
exclusively to one sex.
It's time to
Research from Microsoft has revealed that the number of girls
interested in STEM across Europe,on average,almost doubles
when they have a role model to inspire them.
“I have learned that girls
can do the same as boys.”
"Never limit yourself
because of others' limited
We have to work for girls and young women so that
they can decide freely what they want to be when they
grow up. Because while professional women suffer
from glass ceilings,girls and young women suffer from
“glass corridors” that lead them under the premise of
Europe has to promote social justice in its policies,and it is not
possible to make scientific policy by excluding half of the population.