Den lör 25 juli 2020 kl 01:07 skrev Adam Sobieski <adamsobieski@hotmail.com>:

When exploring how best to add the thematic relation of instrument to the agent-patient pair (e.g. adding “using the bat” to “Bobby hit the ball”), I observed that, for the same input grammatical arguments, there was a set of possible output paraphrases:

 

  1. “Bobby hit the ball using the bat”
  2. “Bobby, using the bat, hit the ball”
  3. “Using the bat, Bobby hit the ball”

 

I wondered: how might we be able to generate each?

 


Since we are generating text for an encyclopedia, do we need to be able to pick which one of them? I think we could have a manual of style that says that method 1. is the one we use. That simplifies the language not only for the reader, but also our task by not trying to be able to create all nuances of spoken language.

/Jan