Whose Language Counts as High Quality? Measuring Language Ideologies in Text Data Selection

By Suchin Gururangan, Dallas Card, Sarah K. Drier, E. K. Gade, Leroy Z. Wang, Zeyu Wang, Luke S. Zettlemoyer, Noah A. Smith. 2022

100 citations recorded by the source corpus.

Language models increasingly rely on massive web crawls for diverse text data. However, these sources are rife with undesirable content. As such, resources like Wikipedia, books, and news often serve as anchors for automatically selecting web text most suitable for language modeling, a process typically referred to as quality filtering. Using a new dataset of U.S. high school newspaper articles—written by students from across the country—we investigate whose language is preferred by the quality filter used for GPT-3. We find that newspapers from larger schools, located in wealthier, educated, and urban zones (ZIP codes) are more likely to be classified as high quality. We also show that this quality measurement is unaligned with other sensible metrics, such as factuality or literary acclaim. We argue that privileging any corpus as high quality entails a language ideology, and more care is needed to construct training corpora for language models, with better transparency and justification for the inclusion or exclusion of various texts.

Subfields: Artificial Intelligence, General Social Sciences, Natural Language Processing, Evaluation & Benchmarks

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 General Social Sciences

In Natural Language Processing

In Evaluation & Benchmarks

Citations in the corpus

Re-contextualizing Fairness in NLP: The Case of India (2022, background): “Further, the risk of unintentionally excluding marginalized communities based on dialect or other linguistic features while filtering data to ensure quality (Dodge et al., 2021; Gururangan et al., 2022) is even higher in the Indian context because of very limited computational representation of marg”
Measuring Data (2022, background): “For example, Gururangan et al. (2022) explore how unexamined deﬁnitions of what constitutes data “quality” may encode ideologies in ways that further suppress marginalized identities.”
Pile of Law: Learning Responsible Data Filtering from the Law and a 256GB Open-Source Legal Dataset (2022, background): “Filtering for toxicity may have unexpected effects on representation in datasets or the bias of downstream outputs [40, 48, 6].”
Prompting PaLM for Translation: Assessing Strategies and Performance (2022, background): “Furthermore, an editorial choice defining high-quality language can potentially reduce quality for groups and topics not typically discussed in this style (Gururangan et al., 2022).”