I’m kinda new to RASA and I’m developing my first chatbot, which is capable of translating words and expressions from Spanish to English.
I’m having trouble getting the word or expression recognized as an entity. So for example, the user input might be something like:
“¿Cómo se dice avión en inglés?” (How do you say plane in English?)
The entity there should be avión (plane), and it could be another word or expression in the Spanish dictionary.
I’ve considered using a lookup table, but that doesn’t seem like a good way to go. Can somebody help me to recognize the entities? Thank you in advance!
I’ve provided about 5 training examples, it’s a small nlu.md file but I’m planing to make it bigger as soon as possible!
Here’s my intent for translation:
## intent:translation
- Cómo se dice [suelo](wordAsked) en inglés
- Como se traduce [lavadora](wordAsked) en inglés
- [Plato](wordAsked) en inglés
- En inglés como se dice [coche](wordAsked)
- Cómo traducir [orca](wordAsked) en inglés
Thing is, user could ask for literally any word to be translated
@AlbertSR 5 examples is really small. You will need a lot more examples. Collecting more examples from users seems like the place to start. Why did you think that a lookup table was not the way to go?
Because they could ask for literally every single word in the spanish dictionary to be translated.
How many examples would be a decent amount in order to recognise entities? Does it matter if some examples have the same sentece structure but different entities? For example: