Detecting entity from a intent not describing it

Rasa Open Source
1

Domain slots: VALUE_MIN: type : float VALUE_MAX: type : float

nlu

intent: start

  • start
  • go

intent: value_min

  • min value is 0-9
  • minimum value is 0-9

intent: value_max

  • max value is 0-9
  • maximum value is 0-9

stories

flow_1

  • start
    • utter_ask_VALUE_MIN
  • value_min{“VALUE_MIN”:“120”}
    • slot{“VALUE_MIN”:“120”}
    • utter_ask_VALUE_MAX
  • value_max{“VALUE_MAX”:“2000”}
    • slot{“VALUE_MAX”:“2000”}

Your input -> go 2020-08-24 13:53:51 DEBUG rasa.core.tracker_store - Recreating tracker for id ‘8ed2a94d33a24c959c42a04acd54cc5c’ 2020-08-24 13:53:51 DEBUG rasa.nlu.classifiers.diet_classifier - There is no trained model for ‘ResponseSelector’: The component is either not trained or didn’t receive enough training data. 2020-08-24 13:53:51 DEBUG rasa.nlu.selectors.response_selector - Adding following selector key to message property: default 2020-08-24 13:53:51 DEBUG rasa.core.processor - Received user message ‘go’ with intent ‘{‘name’: ‘start’, ‘confidence’: 0.7473200559616089}’ and entities ‘[]’ 2020-08-24 13:53:51 DEBUG rasa.core.processor - Logged UserUtterance - tracker now has 27 events. 2020-08-24 13:53:51 DEBUG rasa.core.policies.memoization - Current tracker state [None, None, None, {}, {‘intent_start’: 1.0, ‘prev_action_listen’: 1.0}] 2020-08-24 13:53:51 DEBUG rasa.core.policies.memoization - There is a memorised next action ‘utter_ask_VALUE_MIN’ 2020-08-24 13:53:51 DEBUG rasa.core.policies.ensemble - Predicted next action using policy_0_MemoizationPolicy 2020-08-24 13:53:51 DEBUG rasa.core.processor - Predicted next action ‘utter_ask_VALUE_MIN’ with confidence 1.00. 2020-08-24 13:53:51 DEBUG rasa.core.processor - Action ‘utter_ask_VALUE_MIN’ ended with events ‘[BotUttered(‘what is the min value’, {“elements”: null, “quick_replies”: null, “buttons”: null, “attachment”: null, “image”: null, “custom”: null}, {}, 1598257431.730513)]’. 2020-08-24 13:53:51 DEBUG rasa.core.policies.memoization - Current tracker state [None, None, {}, {‘intent_start’: 1.0, ‘prev_action_listen’: 1.0}, {‘intent_start’: 1.0, ‘prev_utter_ask_VALUE_MIN’: 1.0}] 2020-08-24 13:53:51 DEBUG rasa.core.policies.memoization - There is a memorised next action ‘action_listen’ 2020-08-24 13:53:51 DEBUG rasa.core.policies.mapping_policy - There is no mapped action for the predicted intent, ‘start’. 2020-08-24 13:53:51 DEBUG rasa.core.policies.ensemble - Predicted next action using policy_0_MemoizationPolicy 2020-08-24 13:53:51 DEBUG rasa.core.processor - Predicted next action ‘action_listen’ with confidence 1.00. 2020-08-24 13:53:51 DEBUG rasa.core.processor - Action ‘action_listen’ ended with events ‘[]’. 2020-08-24 13:53:51 DEBUG rasa.core.lock_store - Deleted lock for conversation ‘8ed2a94d33a24c959c42a04acd54cc5c’. what is the min value Your input -> min values is 400 2020-08-24 13:53:59 DEBUG rasa.core.tracker_store - Recreating tracker for id ‘8ed2a94d33a24c959c42a04acd54cc5c’ 2020-08-24 13:53:59 DEBUG rasa.nlu.classifiers.diet_classifier - There is no trained model for ‘ResponseSelector’: The component is either not trained or didn’t receive enough training data. 2020-08-24 13:53:59 DEBUG rasa.nlu.selectors.response_selector - Adding following selector key to message property: default 2020-08-24 13:53:59 DEBUG rasa.core.processor - Received user message ‘min values is 400’ with intent ‘{‘name’: ‘value_min’, ‘confidence’: 0.7056987881660461}’ and entities ‘[{‘entity’: ‘VALUE_MIN’, ‘start’: 14, ‘end’: 17, ‘value’: ‘400’, ‘extractor’: ‘DIETClassifier’}]’ 2020-08-24 13:53:59 DEBUG rasa.core.processor - Current slot values: VALUE_MAX: None VALUE_MIN: 400 2020-08-24 13:53:59 DEBUG rasa.core.processor - Logged UserUtterance - tracker now has 32 events. 2020-08-24 13:54:00 DEBUG rasa.core.policies.memoization - Current tracker state [None, {}, {‘intent_start’: 1.0, ‘prev_action_listen’: 1.0}, {‘intent_start’: 1.0, ‘prev_utter_ask_VALUE_MIN’: 1.0}, {‘prev_action_listen’: 1.0, ‘slot_VALUE_MIN_0’: 1.0, ‘intent_value_min’: 1.0}] 2020-08-24 13:54:00 DEBUG rasa.core.policies.memoization - There is a memorised next action ‘utter_ask_VALUE_MAX’ 2020-08-24 13:54:00 DEBUG rasa.core.policies.ensemble - Predicted next action using policy_0_MemoizationPolicy 2020-08-24 13:54:00 DEBUG rasa.core.processor - Predicted next action ‘utter_ask_VALUE_MAX’ with confidence 1.00. 2020-08-24 13:54:00 DEBUG rasa.core.processor - Action ‘utter_ask_VALUE_MAX’ ended with events ‘[BotUttered(‘what is the max value’, {“elements”: null, “quick_replies”: null, “buttons”: null, “attachment”: null, “image”: null, “custom”: null}, {}, 1598257440.0605109)]’. 2020-08-24 13:54:00 DEBUG rasa.core.policies.memoization - Current tracker state [{}, {‘intent_start’: 1.0, ‘prev_action_listen’: 1.0}, {‘intent_start’: 1.0, ‘prev_utter_ask_VALUE_MIN’: 1.0}, {‘prev_action_listen’: 1.0, ‘slot_VALUE_MIN_0’: 1.0, ‘intent_value_min’: 1.0}, {‘slot_VALUE_MIN_0’: 1.0, ‘intent_value_min’: 1.0, ‘prev_utter_ask_VALUE_MAX’: 1.0}] 2020-08-24 13:54:00 DEBUG rasa.core.policies.memoization - There is a memorised next action ‘action_listen’ 2020-08-24 13:54:00 DEBUG rasa.core.policies.mapping_policy - There is no mapped action for the predicted intent, ‘value_min’. 2020-08-24 13:54:00 DEBUG rasa.core.policies.ensemble - Predicted next action using policy_0_MemoizationPolicy 2020-08-24 13:54:00 DEBUG rasa.core.processor - Predicted next action ‘action_listen’ with confidence 1.00. 2020-08-24 13:54:00 DEBUG rasa.core.processor - Action ‘action_listen’ ended with events ‘[]’. 2020-08-24 13:54:00 DEBUG rasa.core.lock_store - Deleted lock for conversation ‘8ed2a94d33a24c959c42a04acd54cc5c’. what is the max value Your input -> max value is 90000 2020-08-24 13:54:16 DEBUG rasa.core.tracker_store - Recreating tracker for id ‘8ed2a94d33a24c959c42a04acd54cc5c’ 2020-08-24 13:54:16 DEBUG rasa.nlu.classifiers.diet_classifier - There is no trained model for ‘ResponseSelector’: The component is either not trained or didn’t receive enough training data. 2020-08-24 13:54:16 DEBUG rasa.nlu.selectors.response_selector - Adding following selector key to message property: default 2020-08-24 13:54:16 DEBUG rasa.core.processor - Received user message ‘max value is 90000’ with intent ‘{‘name’: ‘value_max’, ‘confidence’: 0.8313025236129761}’ and entities ‘[{‘entity’: ‘VALUE_MIN’, ‘start’: 13, ‘end’: 18, ‘value’: ‘90000’, ‘extractor’: ‘DIETClassifier’}]’ 2020-08-24 13:54:16 DEBUG rasa.core.processor - Current slot values: VALUE_MAX: None VALUE_MIN: 90000 2020-08-24 13:54:16 DEBUG rasa.core.processor - Logged UserUtterance - tracker now has 37 events. 2020-08-24 13:54:16 DEBUG rasa.core.policies.memoization - Current tracker state [{‘intent_start’: 1.0, ‘prev_action_listen’: 1.0}, {‘intent_start’: 1.0, ‘prev_utter_ask_VALUE_MIN’: 1.0}, {‘prev_action_listen’: 1.0, ‘slot_VALUE_MIN_0’: 1.0, ‘intent_value_min’: 1.0}, {‘slot_VALUE_MIN_0’: 1.0, ‘intent_value_min’: 1.0, ‘prev_utter_ask_VALUE_MAX’: 1.0}, {‘intent_value_max’: 1.0, ‘prev_action_listen’: 1.0, ‘slot_VALUE_MIN_0’: 1.0}] 2020-08-24 13:54:16 DEBUG rasa.core.policies.memoization - There is no memorised next action 2020-08-24 13:54:16 DEBUG rasa.core.policies.ensemble - Predicted next action using policy_1_TEDPolicy 2020-08-24 13:54:16 DEBUG rasa.core.processor - Predicted next action ‘action_listen’ with confidence 0.92. 2020-08-24 13:54:16 DEBUG rasa.core.processor - Action ‘action_listen’ ended with events ‘[]’. 2020-08-24 13:54:16 DEBUG rasa.core.lock_store - Deleted lock for conversation ‘8ed2a94d33a24c959c42a04acd54cc5c’. Your input ->

How come for the message ‘max value is 90000’ the intent is identified correct but the entity is not ? How to rectify this ?

2

The NLU section isn’t rendering properly. Could you place it inside a code block (the </> symbol), so we can see the entity training data?

3 
## intent: start
- start
- go

## intent: value_min
- min value is [0-9](VALUE_MIN)
- minimum value is [0-9](VALUE_MIN)

## intent: value_max
- max value is [0-9](VALUE_MAX)
- maximum value is [0-9](VALUE_MAX)
4

Hmm interesting, that looks correct to me.

I’m wondering if this is because VALUE_MIN and VALUE_MAX entities are too similar, so it doesn’t know which one to choose for the number extraction.

I see that there are roles that might help in this case: Entity Extraction

Or you could have a number entity for both, and use the stories to fill the correct slot depending on the classification.