Documenting Computer Vision Datasets: An Invitation to Reflexive Data Practices

Lade...
Vorschaubild
Datum
2021
Herausgeber:innen
Autor:innen
Miceli, Milagros
Yang, Tianling
Naudts, Laurens
Schüßler, Martin
Serbanescu, Diana
Hanna, Alex
Zeitschriftentitel
ISSN der Zeitschrift
Bandtitel
Verlag
Zusammenfassung

In industrial computer vision, discretionary decisions surrounding the production of image training data remain widely undocumented. Recent research taking issue with such opacity has proposed standardized processes for dataset documentation. In this paper, we expand this space of inquiry through fieldwork at two data processing companies and thirty interviews with data workers and computer vision practitioners. We identify four key issues that hinder the documentation of image datasets and the effective retrieval of production contexts. Finally, we propose reflexivity, understood as a collective consideration of social and intellectual factors that lead to praxis, as a necessary precondition for documentation. Reflexive documentation can help to expose the contexts, relations, routines, and power structures that shape data.

Beschreibung
Schlagwörter
datasheets for datasets \ dataset documentation \ reflexivity \ data annotation \ training data \ transparency \ accountability \ audits \ machine learning
Verwandte Ressource
Verwandte Ressource
Zitierform
Miceli, M., Yang, T., Naudts, L., Schüßler, M., Serbanescu, D., & Hanna, A. (2021). Documenting Computer Vision Datasets: An Invitation to Reflexive Data Practices. Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, 161–172. https://doi.org/10.1145/3442188.3445880