[15p-S2_204-1~18] 23.1 Joint Session N "Informatics"

5:45 PM - 6:00 PM JST(8:45 AM - 9:00 AM UTC)

[15p-S2_204-16] Development of LLM Tools to Accelerate Literature Data Collection for Starrydata

〇Yukari Katsura^1,2,3, Tomoya Mato¹, Eiji Koyama¹, Dewi Yana¹, Atsumi Tanaka¹, Yoshihiro Sakamoto³, Masaya Kumagai^3,4 (1.NIMS, 2.Univ. of Tsukuba, 3.RIKEN, 4.SAKURA internet Inc.)

Keywords:

informatics,database,data mining

We developed two LLM-based tools to accelerate data collection for Starrydata. Starrydata Auto-Suggestion for Sample Information is a feature that reads the text in papers and suggests input candidates for pre-designed data items specific to each materials field. Starrydata Auto-Summary GPT is a feature that fully parses open-access paper PDFs uploaded by users and automatically compiles all figures, tables, and sample descriptions appearing in the paper into JSON-formatted data.