They, Them, Theirs: Rewriting with Gender-Neutral English

By Tony Sun, Kellie Webster, Apurva Shah, William Yang Wang, Melvin Johnson. 2021

66 citations recorded by the source corpus.

Responsible development of technology involves applications being inclusive of the diverse set of users they hope to support. An important part of this is understanding the many ways to refer to a person and being able to fluently change between the different forms as needed. We perform a case study on the singular they, a common way to promote gender inclusion in English. We define a re-writing task, create an evaluation benchmark, and show how a model can be trained to produce gender-neutral English with<1% word error rate with no human-labeled data. We discuss the practical applications and ethical considerations of the task, providing direction for future work into inclusive natural language systems.

Subfields: Artificial Intelligence, Natural Language Processing

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

Citations in the corpus

Harms of Gender Exclusivity and Challenges in Non-Binary Representation in Language Technologies (2021, neutral): “This work documents and demonstrates speciﬁc challenges towards making current language modeling techniques inclusive of all genders and re-Pronouns et al. (2021) propose rewriting text containing gen-dered pronouns with they as the substituted pro-noun to obtain more gender-neutral text.”
Societal Biases in Language Generation: Progress and Challenges (2021, background): “Since effective models for NLG tasks are correlated with increasing training data sizes, biases in data collection (e.g., English-centric, drawn from popular Western media) remain a major contributor of biases that manifest in generation.”
First the Worst: Finding Better Gender Translations During Beam Search (2021, unknown): “for monolingual tasks reinflect whole-sentence gender (Habash et al., 2019; Alhafni et al., 2020; Sun et al., 2021).”
Gender Bias in Machine Translation (2021, background): “Besides a few notable exceptions for English NLP tasks (Manzini et al., 2019; Cao and Daumé III, 2020; Sun et al., 2021) and one in MT (Saunders et al.”
Perturbation Augmentation for Fairer NLP (2022, background): “Grammar-based approaches also require ad hoc solutions for phonological alternations ( a banana v. an apple ), and struggle with one-to-many-mappings for pronouns (Sun et al., 2021), often incompletely handling pronoun coreference chains.”
Improving Gender Fairness of Pre-Trained Language Models without Catastrophic Forgetting (2021, background): “One optional term that is already used as a singular pronoun like "they/their" (Darr and Kibbey, 2016; Richards et al., 2016; Sun et al., 2021).”
Bias and Fairness in Large Language Models: A Survey (2023, methodology): “To provide gender-neutral alternatives to sentences with gendered pronouns, several works (Jain et al., 2021; Sun et al., 2021; Vanmassenhove et al., 2021) use a rules-based approach to generate parallel debiased sentences from biased sources, and then train a machine translation model to translate ”
NL-Augmenter: A Framework for Task-Sensitive Natural Language Augmentation (2021, methodology): “This transformation is based on the algorithm proposed by Sun et al. (2021).”
A Survey on Gender Bias in Natural Language Processing (2021, background): “The increased popularity of gender-neutral linguistic forms in natural language presents a challenge to incorporate this social development into the datasets and algorithms [Sun et al. 2021].”