Mapping Research with Neural Networks

This chapter explores the application of neural network algorithms, particularly Word2Vec, in mapping physics research and analyzing the context and temporal distance of scientific papers. The discussion emphasizes a novel approach of treating all works of a scientist as a unified context while examining the potential of tailored embeddings for specific research challenges. Additionally, the chapter examines the validation of a research classification scheme using network science, assessing predictive power, and envisioning future experiments across disciplines.