Harvard University · AI & Machine Learning
HARVARD DATAVERSE
WHAT IT IS ABOUT.
Harvard Dataverse is an open dataset published by Harvard University in the "Research & Machine Learning Data" domain. Available formats: Dataset-specific. License: Record-specific. Update cadence: ongoing deposits.
Suitable for undergraduate and graduate theses, group projects, papers and personal analysis portfolios. Always review the documentation and collection methodology before citing the data: institutions publish detailed methodological notes alongside every series.
Accountra indexes this resource and verifies its free-access requirements; delivery remains the responsibility of Harvard University on the official site.
IN THE SAME VEIN.
CERN / OpenAIREZenodo
Research repository where researchers and organizations publish datasets, software and supplementary materials with persistent identifiers and record-level licences.
University of California, IrvineUCI Machine Learning Repository
University-maintained collection of datasets for machine learning, data mining, teaching and benchmark experimentation.
OpenML FoundationOpenML
Open platform for publishing and discovering machine-learning datasets, tasks, experiments and benchmark results.
Hugging FaceHugging Face Datasets Hub
Dataset hub for machine learning and AI: NLP, computer vision, audio, tabular, time series, and geospatial data, with a built-in viewer and library integration. Free access.