Mining Topic Structure of AI Algorithmic Literature
Hengyi Miao, Xinguo Yu, Hao Wu · 2023
Staying abreast of the latest research findings in any field is a formidable challenge, demanding substantial effort even from experienced researchers. This challenge is further compounded by the rapid proliferation of scientific and technical publications, making it increasingly challenging for individuals to keep up with the expanding knowledge in their respective fields. To address this problem, we propose a model for extracting the topic structure in AI algorithmic literature sets. In this paper, we provide a comprehensive definition of topics in the algorithmic literature and use prompts and large language models to extract topics from individual articles. In addition, we collaborated with domain experts to construct algorithmic topic structure networks based on extracted content. These networks reveal the intricate interconnections between topics, enabling researchers to easily understand current algorithmic advancements in the field. After comparing relevant literature reviews, our study confirms the effectiveness of the defined themes for aiding researchers in comprehending the complex algorithmic relationships in articles concerning AI algorithms.