TODO:
-
can we build a causal graph of the provenance?
-
target publications:
-
impl a semantic random walk from a source KG
-
link entities for lemmas, noun chunks using MediaWiki lookups?
- apply default semantics:
skos:related
- apply default semantics:
-
eval clustering/community detection for GOR?
-
RAG example
-
extend GOR to replicate NodePiece/ULTRA ?
-
reify GOR, then use FastRP to generate embeddings?
-
eval community detection to condense nodes using k-medoids?
-
add conda packaging
-
SPARQL the DBPedia/Wikidata equivs
-
other NER/RE:
-
check out https://github.com./wikipedia2vec/wikipedia2vec
-
link
sense2vec
synonyms; make affordances for UI to annotate synonyms