Automatic Extraction of IDM-Related Information in Scientific Articles and Online Science News Websites

Abstract

Previous studies have made it possible to extract information related to IDM (Inventive Design Method) out of patents. IDM is an ontology-defined method derived from TRIZ. As its mother theory, IDM is primarily based on patent’s observation and aims at finding inventive solutions on the basis of contradictions. In this paper, we present a new approach for extracting knowledge, this time out of other types of science-related documents: scientific papers and science news articles. This approach is based on sets of linguistics features which have been selected and evaluated semi-automatically with techniques of Natural Language Processing as well as Machine Learning.

BibTeX

@inproceedings{nedey:hal-02279756,
  TITLE = {{Automatic Extraction of IDM-Related Information in Scientific Articles and Online Science News Websites}},
  AUTHOR = {N{\'e}dey, Oriane and Souili, Achille and Cavallucci, Denis},
  URL = {https://inria.hal.science/hal-02279756},
  NOTE = {Part 6: TRIZ and Patenting},
  BOOKTITLE = {{IFIP Advances in Information and Communication Technology}},
  ADDRESS = {Strasbourg, France},
  EDITOR = {Denis Cavallucci and Roland De Guio and Sebastian Koziolek},
  PUBLISHER = {{Springer International Publishing}},
  SERIES = {Automated Invention for Smart Industries},
  VOLUME = {AICT-541},
  PAGES = {213-224},
  YEAR = {2018},
  MONTH = Oct,
  DOI = {10.1007/978-3-030-02456-7\_18},
  KEYWORDS = {NLP ; Text mining ; Knowledge extraction ; Machine learning ; Inventive Design ; IDM ; TRIZ},
  PDF = {https://inria.hal.science/hal-02279756v1/file/474537_1_En_18_Chapter.pdf},
  HAL_ID = {hal-02279756},
  HAL_VERSION = {v1},
}