Detecting Independent Pronoun Bias with Partially-Synthetic Data Generation

By Robert Munro, Alex Morrison. 2020

10 citations recorded by the source corpus.

We report that state-of-the-art parsers consistently failed to identify “hers” and “theirs” as pronouns but identified the masculine equivalent “his”. We find that the same biases exist in recent language models like BERT. While some of the bias comes from known sources, like training data with gender imbalances, we find that the bias is _amplified_ in the language models and that linguistic differences between English pronouns that are not inherently biased can become biases in some machine learning models. We introduce a new technique for measuring bias in models, using Bayesian approximations to generate partially-synthetic data from the model itself.

Subfields: Artificial Intelligence, Natural Language Processing, Safety, Robustness & Interpretability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Natural Language Processing

In Safety, Robustness & Interpretability

Citations in the corpus

BBQ: A hand-built bias benchmark for question answering (2021, background): “Though there are other ways of introducing lexical diversity into examples (e.g., Munro and Morrison (2020) mask target words and use an LM to suggest likely words in context), given the extensive validation needed for these templates, other options would give us less control over the exact form of…”
Theories of “Gender” in NLP Bias Research (2022, unknown): “[100] Robert Munro and Alex (Carmen) Morrison.”
This Prompt is Measuring : Evaluating Bias Evaluation in Language Models (2023, background): “…are almost always described as male and female pronouns, though they are widely used by nonbinary individuals 7 (Dev et al., 2021) (an exception is Munro and Morrison (2020), who write of “people who use ‘hers,’ ‘theirs’ and ‘themself’ to align their current social gender(s) with their pronouns’…”
On Measures of Biases and Harms in NLP (2021, background): “For part-of-speech tagging, Munro and Morrison (2020) and Garimella et al. (2019) find that state-of-the-art parsers perform differently across genders, failing to identify “hers” and “theirs” as pronouns but not “his”.”
What do Bias Measures Measure? (2021, background): “For part-of-speech tagging, Munro and Morrison (2020) and Garimella et al. (2019) ﬁnd that state-of-the-art parsers perform differently across genders, failing to identify “hers” and “theirs” as pronouns but not “his”.”