Hi, any idea/code how to automate entity synonym training in RASA NLU?
By default, you have to make for every synonym separate training sentence. Would not be great to automate this? When polishing NLU performance, you would just add more synonyms to existing entities…
Thanks for sharing
Use a lookup table.
Thanks! Lookup tables are great for better entity extraction.
But do not solve standardization of entity value. For languages with complex morphology (e.g. CZ/PL/SK) it is often necessary to put entity into standard form (nominativ).
I have the same question, hi from 2022. Still no solution?