✳flâneur — a map of the web's best reading

The Data Cards Playbook - Data Cards Playbook

sites.research.google · 5,422 words · saved by 1 readers

Data Cards are structured summaries of essential facts about various aspects of ML datasets needed by stakeholders across a project's lifecycle for responsible AI development. This Data Card template captures 15 themes that we frequently look for when making decisions — many of which are not traditionally captured in technical dataset documentation. Click on a theme below to see it in the Data Card and learn more: Publishers, Dataset Owners, Funding Sources Sensitivity of Data, Dataset Version and Maintenance Motivations, Intended Use Access, Retention, Wipeout and Deletion Collection, Collection Criteria, Relationship to Source, Version and Maintenance Use with Other Data, Forking & Sampling, Use in ML or AI Systems Synopsis, Breakdown of Transformations Human Annotators Description of Human Validators Concepts and Definitions referenced in this Data Card Write a short summary describing your dataset (limit 200 words). Include information about the content and topic of the data, sourc

Dataset Name (Acronym) Write a short summary describing your dataset (limit 200 words). Include information about the content and topic of the data, sources and motivations for the dataset, benefits and the problems or use cases it is suitable for. Dataset Link Dataset Link Data Card Author(s) Name, Team: (Owner / Contributor / Manager) Name, Team: (Owner / Contributor / Manager) Name, Team: (Owner / Contributor / Manager) Authorship Publishers, Dataset Owners, Funding Sources Publishers Publishing Organization(s) Organization Name Industry Type(s) Corporate - Tech Corporate - Non-Tech (please

Explore this link on the map →

saved by

related reading