Abstract
Embedding spaces define notions of semantic similarity and distance. We study whether those embeddings reflect physical measurements of mass, distance, time and volume, which admit a unique, objective notion of semantic equivalence and distance. We find that physical measurement is only weakly modeled in the embedding space, and that instead quite peculiar measurement patterns can be observed. Further analysis indicates that embedding representations of physical measurements are strongly influenced by superficial string similarity, and recalibration of similarity does not substantially improve the alignment.
Keywords
Subject
Publication details
- Journal
- Not available
- Open access
- Green open access
Cite this article
APA 7
Opitz, J., & Michail, A. (2026). Embedding Models Measure in Peculiar Ways. https://omanscience.com/en/articles/embedding-models-measure-in-peculiar-ways
MLA 9
Opitz, Juri, and Andrianos Michail. "Embedding Models Measure in Peculiar Ways." https://omanscience.com/en/articles/embedding-models-measure-in-peculiar-ways.
Chicago (author–date)
Opitz, Juri, and Andrianos Michail. 2026. "Embedding Models Measure in Peculiar Ways." https://omanscience.com/en/articles/embedding-models-measure-in-peculiar-ways.
Harvard
Opitz, J. and Michail, A. (2026) 'Embedding Models Measure in Peculiar Ways', Available at: https://omanscience.com/en/articles/embedding-models-measure-in-peculiar-ways.
Vancouver
Opitz J, Michail A. Embedding Models Measure in Peculiar Ways. https://omanscience.com/en/articles/embedding-models-measure-in-peculiar-ways
IEEE
J. Opitz, and A. Michail, "Embedding Models Measure in Peculiar Ways," https://omanscience.com/en/articles/embedding-models-measure-in-peculiar-ways.