Toshiba AI Technology Catalog

  • Language media analysis

Extraction of similar expressions

Quantifying semantic similarity for combined expressions (n-grams) of multiple words.

  • By preparing networks for each n-gram length and by forming connections between hidden and output layers, weights are updated in consideration of differences in each n-gram length.
  • By accurately identifying semantic similarity for n-grams, it becomes possible to efficiently find the desired document from among large-scale document data.


  • Text search/classification system

Benchmarks, strengths, and track record

  • Accurately estimates semantic similarity for combined expressions of multiple words compared to conventional methods. (Association for Natural Language Processing, 25th Annual Meeting)


Please include the title “Toshiba AI Technology Catalog: Extraction of synonymous expressions” or the URL in the inquiry text.
Please note that because this technology is currently the subject of R&D activities, immediate responses to inquiries may not be possible.