Network Text Analysis Framework for Mapping Research Trends: A Neural Architecture Search Case Study

Abstrakt

Understanding the evolution of research trends is critical for navigating rapidly developing scientific literature. Large Language Models (LLMs) offer powerful tools for analysing scientific texts, enabling the extraction of key concepts and the construction of semantic networks. These capabilities can support the study of emerging ideas and research trends through graph-based representations. In this paper, we present a network-based text analysis framework designed to map the evolution of scientific knowledge. Our goal is to extract conceptual structures from research papers and construct graphs that represent both the occurrence of terms and their interrelationships. The integration of temporal information allows us to track the emergence and transformation of research themes. We demonstrate this framework using a case study on Neural Architecture Search (NAS) field, a fast-growing subfield in machine learning focused on the automated design of neural networks. Using data from ArXiv combined with metadata and citation records from OpenAlex, we construct and analyse graphs of keywords and articles. This allows us to reveal the dynamics of the NAS research landscape and highlight methodological trends and conceptual shifts.