Abstract
This position paper takes Niels’ Lund concept of doceme and applies it to an analysis of a contemporary information retrieval system for Wikipedia, constructed using the Vector Database system “Upstash Vector.” It begins with a theoretical analysis of the capacity for machine learning models to comprehend language, situating its position within that wide debate. It proceeds to discuss issues between the use of the term “semantic” in an data science vs. information science context. Finally, it applies theories from Lund’s doceme and Ron Day’s Documentarity to an analysis of the Upstash Vector dataset located on hugging face, closing with a qualitative reflection on what such systems indicate for the future of documentation.
Recommended Citation
Kausch, John
(2025)
"Docemes and Document Embeddings,"
Proceedings from the Document Academy: Vol. 12
:
Iss.
2
, Article 3.
DOI: https://doi.org/10.35492/docam/12/2/3
Available at:
https://ideaexchange.uakron.edu/docam/vol12/iss2/3
Digital Object Identifier (DOI)
10.35492/docam/12/2/3