Ben Hutchinson, Andrew Smart, Alex Hanna, Emily Denton, Christina Greer, Oddur Kjartansson, Parker Barnes, & Margaret Mitchell (2021)
Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, 560-575.
DOI: https://doi.org/10.1145/3442188.3445918
Abstract. Hutchinson et al.'s argument for treating ML datasets as maintained infrastructure with versioning, documentation, and audit trails — borrowing accountability practices from software engineering. Influential in dataset-card and model-card design.
Tags: ai-usability accountability datasets