Corpus ID: 1704893

Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings

@inproceedings{Bolukbasi2016ManIT,
  title={Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings},
  author={Tolga Bolukbasi and Kai-Wei Chang and James Y. Zou and Venkatesh Saligrama and A. Kalai},
  booktitle={NIPS},
  year={2016}
}
The blind application of machine learning runs the risk of amplifying biases present in data. [...] Key Method Geometrically, gender bias is first shown to be captured by a direction in the word embedding. Second, gender neutral words are shown to be linearly separable from gender definition words in the word embedding.Expand
View PDF on arXiv
1,269 Citations
Highly Influential Citations
192
Background Citations
755
Methods Citations
239
Results Citations
35

Supplemental Content

Github Repo
Via Papers with Code
uzairakbar/info-retrieval
Information Retrieval in High Dimensional Data (class deliverables)
28 Stars • 9 Forks
2:42
Video
NIPS16 - Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings
Tolga Bolukbasi
22 November 2016

1,269 Citations

Citation Type

Citation Type

Fair Is Better than Sensational: Man Is to Doctor as Woman Is to Doctor
TLDR
A series of clarifications are provided that should put well-known, and potentially new analogies into the right perspective, which might have yielded a distorted picture of bias in word embeddings. Expand
Adversarial Approaches to Debiasing Word Embeddings
In recent years, word embeddings have been ever more important in the world of natural language processing: techniques such as GloVe and Word2Vec have successfully mapped words to n-dimensional… Expand
Understanding the Origins of Bias in Word Embeddings
TLDR
Given a word embedding trained on a corpus, this work develops a technique for understanding the origins of bias in word embeddings that identifies how perturbing the corpus will affect the bias of the resulting embedding. Expand
What are the Biases in My Word Embedding?
TLDR
An algorithm for enumerating biases in word embeddings that outputs a number of Word Embedding Association Tests (WEATs) that capture various biases present in the data, which makes it easier to identify biases against intersectional groups, which depend on combinations of sensitive features. Expand
Bias in word embeddings
TLDR
A new technique for bias detection for gendered languages is developed and used to compare bias in embeddings trained on Wikipedia and on political social media data, and it is proved that existing biases are transferred to further machine learning models. Expand
Nurse is Closer to Woman than Surgeon? Mitigating Gender-Biased Proximities in Word Embeddings
TLDR
RAN-Debias is proposed, a novel gender debiasing methodology that not only eliminates the bias present in a word vector but also alters the spatial distribution of its neighboring vectors, achieving a bias-free setting while maintaining minimal semantic offset. Expand
A Source-Criticism Debiasing Method for GloVe Embeddings
TLDR
A simple yet effective method for debiasing GloVe word embeddings which works by incorporating explicit information about training set bias rather than removing biased data outright, and reduces the effect size on Word Embedding Association Test (WEAT) sets without sacrificing training data or TOP-1 performance. Expand
Humpty Dumpty: Controlling Word Meanings via Corpus Poisoning
TLDR
This work develops an explicit expression over corpus features that serves as a proxy for distance between words and establishes a causative relationship between its values and embedding distances, and shows how the attacker can generate linguistically likely corpus modifications, thus fooling defenses that attempt to filter implausible sentences from the corpus using a language model. Expand
"Thy algorithm shalt not bear false witness": An Evaluation of Multiclass Debiasing Methods on Word Embeddings
TLDR
By investigating the religious bias removal on three widely used word embeddings, namely: Word2Vec, GloVe, and ConceptNet, it is shown that the preferred method is ConceptorDebiasing, which manages to decrease the measured religious bias on average by 82,42, 96,78% and 54,76% for the three word embedding sets respectively. Expand
Determining Bias in Machine Translation with Deep Learning Techniques
TLDR
A set of debiased pre-trained word embeddings, vector representation of words, in the Transformer, a neural network architecture for machine translation, are used to study how debiasing affects the translation models. Expand
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 55 REFERENCES
GloVe: Global Vectors for Word Representation
TLDR
A new global logbilinear regression model that combines the advantages of the two major model families in the literature: global matrix factorization and local context window methods and produces a vector space with meaningful substructure. Expand
Linguistic Regularities in Continuous Space Word Representations
TLDR
The vector-space word representations that are implicitly learned by the input-layer weights are found to be surprisingly good at capturing syntactic and semantic regularities in language, and that each relationship is characterized by a relation-specific vector offset. Expand
Linguistic Regularities in Sparse and Explicit Word Representations
TLDR
It is demonstrated that analogy recovery is not restricted to neural word embeddings, and that a similar amount of relational similarities can be recovered from traditional distributional word representations. Expand
Deep Recursive Neural Networks for Compositionality in Language
TLDR
The results show that deep RNNs outperform associated shallow counterparts that employ the same number of parameters and outperforms previous baselines on the sentiment analysis task, including a multiplicative RNN variant as well as the recently introduced paragraph vectors. Expand
Distributed Representations of Words and Phrases and their Compositionality
TLDR
This paper presents a simple method for finding phrases in text, and shows that learning good vector representations for millions of phrases is possible and describes a simple alternative to the hierarchical softmax called negative sampling. Expand
Improving Document Ranking with Dual Word Embeddings
TLDR
This paper investigates the popular neural word embedding method Word2vec as a source of evidence in document ranking and proposes the proposed Dual Embedding Space Model (DESM), which provides evidence that a document is about a query term. Expand
Automated Experiments on Ad Privacy Settings
TLDR
AdFisher, an automated tool that explores how user behaviors, Google's ads, and Ad Settings interact, finds that the Ad Settings was opaque about some features of a user’s profile, that it does provide some choice on advertisements, and that these choices can lead to seemingly discriminatory ads. Expand
Certifying and Removing Disparate Impact
TLDR
This work links disparate impact to a measure of classification accuracy that while known, has received relatively little attention and proposes a test for disparate impact based on how well the protected class can be predicted from the other attributes. Expand
Semi-supervised Question Retrieval with Gated Convolutions
TLDR
This paper designs a recurrent and convolutional model (gated convolution) to effectively map questions to their semantic representations and demonstrates that the model yields substantial gains over a standard IR baseline and various neural network architectures (including CNNs, LSTMs and GRUs). Expand
It's a Man's Wikipedia? Assessing Gender Inequality in an Online Encyclopedia
TLDR
This paper presents and applies a computational method for assessing gender bias on Wikipedia along multiple dimensions and finds that while women on Wikipedia are covered and featured well in many Wikipedia language editions, the way women are portrayed starkly differs from the way men are portrayed. Expand
...
1
2
3
4
5
...