Incorrect intent recognition for similar words in rasa nlu

Rasa Open Source
1

When the user asks out_of_scope question which contains some keyword(helping words to construct the sentence) which is used in defined intent then it chooses the defined intent(I have not used any entity approach).

config.yml

# Configuration for Rasa NLU.
# https://rasa.com/docs/rasa/nlu/components/
language: en
pipeline: supervised_embeddings

# Configuration for Rasa Core.
# https://rasa.com/docs/rasa/core/policies/
policies:
  - name: MemoizationPolicy
    max_history: 5
  - name: KerasPolicy
    epochs: 400
    batch_size: 100
    validation_split: 0.2
    max_history: 5
  - name: MappingPolicy
  - name: "FallbackPolicy"
    nlu_threshold: 0.7
    core_threshold: 0.5
    fallback_action_name: "action_default_fallback"

Below is my intent

intent: ask_faq_how_many _vegetarian_restaurants_are_there_nearby
- how many vegetarian restaurants are there nearby
- vegetarian restaurants near by
- please tell me how many vegetarian restaurants are there

If the user asks :

  1. “what vegetarian food option do i have”?
  2. “do you like vegetarian food”?
  3. “vegetarian food”

then the nlu pick ask_faq_how_many _vegetarian_restaurants_are_there_nearby as intent .

Above 3 questions are not related to any intent and user can have many other chitchats using above keyword and training it all keyword might take a long time. Is there any way we can tell nlu to not to pick intent just based on a few keywords?

2

Can anybody please help

3

In order for an algorithm to understand the difference, you need to provide another intent so that it learns to predict ask_faq_how_many only if “how many” words are present