Project and method
About MatSci-SAM
MatSci-SAM (Semantic Alignment Metadata) is a community metadata dictionary for developing shared materials science terminology through expert contributions, discussion, and human-in-the-group AI.
Developed by the Metadata Research Center at Drexel University, the project records definitions as structured metadata and preserves how people and named models contribute to their refinement. The resulting vocabulary supports clearer data description and the reuse of materials science results.
Looking for instructions? Open the MatSci-SAM user guide.
Community method
How a definition develops
A contribution is available for review when it is submitted. A small set of explicit actions then builds a visible history around it.
- 1
New term
Add a missing term and its first definition. Optionally, prompt a language model for an editable draft.
- 2
Revise or replace
Explain what is wrong to draft a revision, or propose a separate replacement for people to compare and vote on.
- 3
Comment or illustrate
Post a comment as written, or add another example of use. Neither action changes the definition.
- 4
Preserve
Retain attribution, versions, decisions, and model involvement in a traceable record.
AI assistance is confined to New term and Suggest a revision. Nothing publishes automatically, and an accepted draft credits both the contributor and the named model. Examples remain independent so each definition can keep more than one.
Open metadata
Vocabulary content and curation history are machine-readable
SKOS represents terms, definitions, and examples. Dublin Core records contributors and dates. PROV-O expresses revisions, comments, votes, and AI-assisted refinement activities. Together, these formats support FAIR access and reuse.
- SKOS
- Terms and definitions
- Dublin Core
- Attribution and dates
- PROV-O
- Curation history
- FAIR
- Access and reuse
skos:prefLabel "martensite"@en
skos:definition "Community definition..."@en
dcterms:contributor "Contributor"
prov:wasGeneratedBy refinement-roundThe vocabulary record describes the published content. The provenance record describes the activities that produced it.