Quick Question • A doctor is walking down the street with a boy. The boy is the doctor’s son, but the doctor is not the boy’s father. How is that possible? GENDER BIAS IN WORD EMBEDDINGS 1
Simple Answer • The doctor is the boy’s mother… GENDER BIAS IN WORD EMBEDDINGS 2
Gender Bias in Word Embeddings HILA GONEN AND YOAV GOLDBERG BAR ILAN UNIVERSITY WIDS TLV 5/3/19 ACCEPTED TO NAACL 2019
Outline • Background • Gender Bias • Word embeddings • Current debiasing methods • Post-processing (Bolukbasi et al.) • During training (Zhao et al.) • Analyzing debiased embeddings • Conclusion GENDER BIAS IN WORD EMBEDDINGS 4
What do we mean by gender bias? GENDER BIAS IN WORD EMBEDDINGS 5
What do we mean by gender bias? GENDER BIAS IN WORD EMBEDDINGS 5
What do we mean by gender bias? GENDER BIAS IN WORD EMBEDDINGS 6
What do we mean by gender bias? GENDER BIAS IN WORD EMBEDDINGS 6
What do we mean by gender bias? (Zhao et al.,NAACL, 2018) GENDER BIAS IN WORD EMBEDDINGS 7
What do we mean by gender bias? (Hendricks et al., 2018) GENDER BIAS IN WORD EMBEDDINGS 8
Word Embeddings • We will focus on gender bias in word embeddings GENDER BIAS IN WORD EMBEDDINGS 9
Word Embeddings • We will focus on gender bias in word embeddings What are word embeddings? GENDER BIAS IN WORD EMBEDDINGS 9
Word Embeddings • Each word in the vocabulary is represented by a low dimensional vector (~ ) • All words are embedded into the same space • Similar words have similar vectors • (= their vectors are close to each other in the vector space) • Word embeddings are successfully used for various NLP applications GENDER BIAS IN WORD EMBEDDINGS 10
Training Word Embeddings • Learned from raw data • The Distributional Hypothesis: • words that occur in the same contexts tend to have similar meanings (Harris, 1954) • “You shall know a word by the company it keeps” (Firth, 1957) GENDER BIAS IN WORD EMBEDDINGS 11
LANGUAGE MODELING FOR CODE SWITCHING 15 12
LANGUAGE MODELING FOR CODE SWITCHING 16 12
Word Embeddings • TopK lists: dog (Mikolov et al. 2013) GENDER BIAS IN WORD EMBEDDINGS 13
Word Embeddings • TopK lists: food (Mikolov et al. 2013) GENDER BIAS IN WORD EMBEDDINGS 13
Word Embeddings • TopK lists: nurse (Mikolov et al. 2013) GENDER BIAS IN WORD EMBEDDINGS 13
Word Embeddings • TopK lists: nurse (Mikolov et al. 2013) GENDER BIAS IN WORD EMBEDDINGS 13
Bias in Word Embeddings GENDER BIAS IN WORD EMBEDDINGS 14
Bias in Word Embeddings • Caliskan et al. replicate a spectrum of known biases from the literature using word embeddings • Show that text corpora contain several types of biases: • morally neutral as toward insects or flowers • problematic as toward race or gender • veridical, reflecting the distribution of gender with respect to careers or first names • Introduce methods for identifying these biases GENDER BIAS IN WORD EMBEDDINGS 15
Bias in Word Embeddings GENDER BIAS IN WORD EMBEDDINGS 16
Bias in Word Embeddings Concepts 1 Concepts 2 Attributes 1 Attributes 2 Flowers : Insects : Pleasant : Unpleasant : buttercup, daisy, lily ant, caterpillar, flea freedom, health, love abuse, crash, filth European American names : African American names : Pleasant : Unpleasant : Brad, Brendan Darnell, Lakisha joy, love, peace agony, terrible Male attributes : Female attributes : Math words : Arts Words : male, man, boy female, woman, girl math, algebra, geometry poetry, art, dance GENDER BIAS IN WORD EMBEDDINGS 17
Definition of Gender Bias in Word Embeddings (NIPS, 2016) GENDER BIAS IN WORD EMBEDDINGS 18
Definition of Gender Bias in Word Embeddings • We check how similar a word is to “he” and “she” (cosine similarity) • Note that we care about the difference between the two • This is the projection on the direction of “he – she”* * This is the gender direction, can be computed using several pairs together (e.g. man-woman, brother-sister) GENDER BIAS IN WORD EMBEDDINGS 19
Definition of Gender Bias in Word Embeddings • bias(consultant) = -0.0023 • bias(nurse) = -0.2471 • bias(captain) = 0.1521 • bias(table) = -0.0003 Zhao et al. GENDER BIAS IN WORD EMBEDDINGS 20
Reduce Bias after Training • Bolukbasi et al. suggest to remove bias after training: • Define a gender direction • Define inherently neutral words (nurse as opposed to mother) • Zero the projection of all neutral words on the gender direction: 𝑥 Projection of on gender direction • The bias of all neutral words is now zero by definition • We will address these embeddings as HARD-DEBIASED GENDER BIAS IN WORD EMBEDDINGS 21
Reduce bias during training (EMNLP, 2018) GENDER BIAS IN WORD EMBEDDINGS 22
Reduce bias during training • Zhao et al. suggest to reduce bias during training: • Train word embeddings using GloVe (Pennington et al., 2014) • Alter the loss to encourage the gender information to concentrate in the last coordinate • To ignore gender information – simply remove the last coordinate GENDER BIAS IN WORD EMBEDDINGS 23
Reduce bias during training • Details: • Use two groups of male/female seed words, and encourage words from different groups to differ in their last coordinate • Encourage the representation of neutral-gender words (excluding the last coordinate) to be orthogonal to the gender direction • We will address these embeddings as GN-GLOVE GENDER BIAS IN WORD EMBEDDINGS 24
Some Results • Bolukbasi et al.: • Bias of all inherently-neutral words is zero by definition • Generated analogies are less stereotyped • Zhao et al.: • Decrease bias in co-reference resolution GENDER BIAS IN WORD EMBEDDINGS 25
Problem solved? GENDER BIAS IN WORD EMBEDDINGS 26
Problem solved? • Not so fast… GENDER BIAS IN WORD EMBEDDINGS 26
Clustering male- and female- biased words • We take the most biased words in the vocabulary according to the original bias (500 male-biased and 500 female-biased) • We cluster them into two clusters using K-means • The clusters align with gender with accuracy of: • 92.5% compared to 99.99% (HARD-DEBIASED) • 85.6% compared to 100% (GN-GLOVE) GENDER BIAS IN WORD EMBEDDINGS 27
Clustering male- and female- biased words HARD-DEBIASED GN-GLOVE GENDER BIAS IN WORD EMBEDDINGS 28
Bias-by-neighbors • Bias is still manifested by the word being close to socially-marked feminine words • A new mechanism for measuring bias: • The percentage of male/female socially-biased words among the k nearest neighbors of the target word • Pearson correlation with bias-by-projection: • 0.686 compared to 0.741 (HARD-DEBIASED) • 0.736 compared to 0.773 (GN-GLOVE) GENDER BIAS IN WORD EMBEDDINGS 29
Professions • We take a predefined list of professions • We show correlation between the bias-by-projection and bias-by- neighbors, before and after debiasing GENDER BIAS IN WORD EMBEDDINGS 30
Professions HARD-DEBIASED GN-GLOVE GENDER BIAS IN WORD EMBEDDINGS 31
Association with stereotyped words • We evaluate the association between female/male names and female/male stereotyped words (experiments taken from Caliskan et al.) Female-associated Male-associated names Amy, Joan, Lisa John, Paul, Mike family vs. carrer home, parents, children executive, management, professional arts vs. math poetry, art, dance math, algebra, geometry arts vs. science dance, literature, novel science, technology, physics • All the associations have very small p-values GENDER BIAS IN WORD EMBEDDINGS 32
Classifying to gender • Can a classifier learn to generalize from some gendered words to others based only on their representations? train 1000 SVM 5000 most biased words 4000 test GENDER BIAS IN WORD EMBEDDINGS 33
Classifying to gender • Results: GENDER BIAS IN WORD EMBEDDINGS 34
Conclusion • Word embeddings exhibit gender bias • Debiasing is hard! • Social gender bias is picked up from the data by the models • A lot of the bias information is still recoverable (even when the bias is low/zero according to the definition usually used) • The way we define the bias is important, and needs to be reconsidered when trying to solve the problem GENDER BIAS IN WORD EMBEDDINGS 35
Questions? GENDER BIAS IN WORD EMBEDDINGS 36
Thank you! GENDER BIAS IN WORD EMBEDDINGS 37
Recommend
More recommend
