i'm trying to make a slot in amazon lex to detect places in the utterance (this would include address, zip code, land marks, etc), which is to be used for a navigation app. i made a custom slot called "Place", which is supposed to generate a response from the slot when the user says "take me to [place]". however, most of the time, it doesn't understand the utterance. it works when i list down names of places in the possible values (e.g. Tokyo Station, Shibuya, etc), but won't recognize values other than the ones listed. is there way for me to make lex dynamically recognize all of the places in a region, say, Japan?
You can try the slots like AMAZON.US_CITY
, AMAZON.GB_CITY
, AMAZON.EUROPE_CITY
, AMAZON.DE_CITY
, AMAZON.AT_CITY
, etc from the built-in slots.
They can pretty much identify the cities all around the world.
I have checked AMAZON.US_CITY
and it is able to recognize the cities of Japan.
See screenshots below.
In Alexa you can extend the built-in slots as well, but I'm not sure if its possible in Lex.
But AMAZON.US_CITY
will solve your problem.
Hope it helps.