… is essentially a form of engineering practice (Section 3). Building on this, we show how
software engineering and infrastructure practices provide lessons for dataset accountability (…

… , datasets often exhibit past trends of social unfair practices, which machine learning … To
completely accept the advancement of AI research procedures towards socially acceptable …

… transparency which supports decision-making and accountability. The framework … engineering
nature of dataset development to draw on best practices from the software development …

… -parameters, and dataset refinement. Similar experimental … , while at the same time accounting
for the wide variety of ML … they work with machine learning and offer guidance towards …

… We read Bovens as gesturing toward four key parameters in any relational accountability …
Towards accountability for machine learning datasets: Practices from software engineering and …

… (ML) models, like violations of privacy, fairness, and accountability. … machine learning from
traditional software and statistical methods; for the purposes of this study, “machine learning” …

… for ethical AI practices to ensure fairness and accountability. We first … can work towards creating
more equitable AI technologies. … This helps in creating balanced datasets that capture the …

… The dependency of GenAI on extensive datasets not only … sources, dataset characteristics,
collection methods, preprocessing … , responsibility, and robustness in machine learning. In …

… Towards accountability for machine learning datasets: Practices from software
engineering and infrastructure. In Proceedings of the 2021 ACM Conference on Fairness, Ac- …

… For each feature in the dataset, Themis generates test suites … We believe that the first step
toward the proper treatment of … The individual and collaborative responsibilities that each role …