Nitrogen is a prototype natural language generation system that combines symbolic rules with linguistic information gathered statistically from large online text corpora.

It is used in the Gazelle machine translation system to turn meaning representations into English. These representations are made up of concepts and relations in the browsable WordNet-based SENSUS knowledge base, which contains upwards of 70,000 conceptual entities. Click here for sample representations and descriptions of semantic relations used by Nitrogen.

Try Nitrogen

  • choose a meaning representation from this list

    (H / |detest|
       :agent they
       :patient (H2 / |hypocrisy<falseness|))
    (A / "acquit"
       :agent (J / "jury")
       :patient she)
    (A / |obligatory<necessary|
       :domain (A2 / |adapt<vary|
                  :patient (M / |method<knowhow|)
                  :destination (C / |circumstance<status|)))
    (A / |possible<latent|
       :domain (A2 / |come with|
                  :agent (U / |unemployment|
                            :mod (H / |high>higher|))
                  :patient (R / |go up<grow|
                              :agent (P / |terms<cost|))))
    (A / |possible>workable|
       :domain (A2 / |sell<cozen|
                   :agent i
                   :patient (T / |reliance,trust|
                               :gpi they))
       :polarity negative)

  • type in your own