What are the biases in my word embedding?
This addresses the issue of harmful biases in widely used word embeddings, which can perpetuate discrimination in AI applications, though it is incremental in automating bias detection.
The paper tackles the problem of identifying biases in word embeddings by presenting an unsupervised algorithm that enumerates offensive associations related to sensitive features like race and gender, exposing biases even in supposedly 'debiased' embeddings. It demonstrates the algorithm's utility on publicly available embeddings and evaluates its output through crowdsourcing.
This paper presents an algorithm for enumerating biases in word embeddings. The algorithm exposes a large number of offensive associations related to sensitive features such as race and gender on publicly available embeddings, including a supposedly "debiased" embedding. These biases are concerning in light of the widespread use of word embeddings. The associations are identified by geometric patterns in word embeddings that run parallel between people's names and common lower-case tokens. The algorithm is highly unsupervised: it does not even require the sensitive features to be pre-specified. This is desirable because: (a) many forms of discrimination--such as racial discrimination--are linked to social constructs that may vary depending on the context, rather than to categories with fixed definitions; and (b) it makes it easier to identify biases against intersectional groups, which depend on combinations of sensitive features. The inputs to our algorithm are a list of target tokens, e.g. names, and a word embedding. It outputs a number of Word Embedding Association Tests (WEATs) that capture various biases present in the data. We illustrate the utility of our approach on publicly available word embeddings and lists of names, and evaluate its output using crowdsourcing. We also show how removing names may not remove potential proxy bias.