The-AI-Alliance/open-trusted-data-initiative
Clarify "canonical" labels to support
Offen
#108 geöffnet am 04.03.2025
catalogenhancementgood first issuehelp wantedopen data spec
Repository-Metriken
- Stars
- (16 Sterne)
- PR-Merge-Metriken
- (PR-Metriken ausstehend)
Beschreibung
I added two links in the references to general categories used in advertising and NLP. They are hierarchical with the lowest layers probably too fine-grained, but the higher layers could be good categories to provide canned labels for users searching for domain-specific datasets.
Other important labels include: synthetic, training, tuning, materials, healthcare, time series, science, math, tech, audio, video, multimedia, language, particular languages, etc.