New preprint: Trajectories of Change — Approaches for Tracking Knowledge Evolution

Malte Vogl | Dec 31, 2024 min read
Stream graph of clusters from 1950 to 2005, growing wider over time and colored by cluster number.

Distinguishing gradual change from the emergence of genuinely new directions is a central challenge in tracking how research fields evolve. This preprint explores local versus global evolution of knowledge systems through the socio-epistemic networks (SEN) framework, pairing an information-theoretic measure of semantic shift with density estimation over text embeddings to detect and characterize change in a corpus of scientific texts.

Read the paper

  • Authors: Raphael Schlattmann and Malte Vogl
  • Preprint: arXiv:2501.00391
  • Featured image: figure from the preprint