Time-dependent Development of Scientific Discourse: A Novel Approach Using UMAP and Word Embeddings
Keywords: visualization, umap, dimension reduction, history of science, word embeddings
TL;DR: Projecting word embeddings into a low-dimensional space in order to visualize the development of scientific discourse and identify trends and notable events.
Abstract: This study presents a method for visualizing the time-dependent development of a scientific discipline using UMAP (Uniform Manifold Approximation and Projection) and text embeddings. This study demonstrates how the evolution of research interests and topics in a specific field can be mapped over time by encoding the abstracts of scholarly articles into a high-dimensional space and then projecting them into a 3D space. This computational approach converts the history of discourse into a point-cloud that can be further studied as a manifold and as a time series, which leads to new insights into the dynamics of scholarly discourse and the emergence and disappearance of research themes.
Primary Area: learning on graphs and other geometries & topologies
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 7602
Loading