Presentation Information

[24a-12C-3]Entity Recognition/Deduplication of Spelling Inconsistency on Materials Property Names using ChatGPT

〇Michiko Yoshitake1, Takahiro Nagata1 (1.NIMS)

Keywords:

names of material properties,natural language processing,spelling inconsistency

We have been constructing knowlege graph on the relationship between different material properties by extracting information from textbooks. In the construction process, entity recognition/deduplication of spelling inconsistency has been always a problem, such as 'band gap', 'gap energy', and 'energy gap'. We tried to tackle to this problem using ChatGPT.