I created an insurance chatbot. It is working good. I have used a dialogflow small talk . It is giving the default fall back action for all the intents. If I use small talk alone It is working. But combined with insurance bot stories, that problem occurs.Here it is the evaluated core model.failed_stories.md (21.9 KB)
Hi @Tanja, Thanks for the response. Here it is the log details.
Bot loaded. Type a message and press enter (use '/stop' to exit):
Your input -> Hi
2019-11-08 15:00:57 DEBUG rasa.core.agent - Created a new lock for conversation 'default'
2019-11-08 15:00:57 DEBUG rasa.core.tracker_store - Creating a new tracker for id 'default'.
2019-11-08 15:00:57 DEBUG rasa.core.processor - Received user message 'Hi' with intent '{'name': 'smalltalk.greetings.hello', 'confidence': 0.9539153575897217}' and entities '[]'
2019-11-08 15:00:57 DEBUG rasa.core.processor - Logged UserUtterance - tracker now has 2 events
2019-11-08 15:00:57 DEBUG rasa.core.policies.fallback - NLU confidence threshold met, confidence of fallback action set to core threshold (0.3).
2019-11-08 15:00:57 DEBUG rasa.core.policies.form_policy - There is no active form
2019-11-08 15:00:57 DEBUG rasa.core.policies.ensemble - Predicted next action using policy_1_FallbackPolicy
2019-11-08 15:00:57 DEBUG rasa.core.processor - Predicted next action 'utter_default' with confidence 0.30.
2019-11-08 15:00:57 DEBUG rasa.core.processor - Action 'utter_default' ended with events '['BotUttered(text: I was unable to understand that, please try asking in a different way., data: {"elements": null, "quick_replies": null, "buttons": null, "attachment": null, "image": null, "custom": null}, metadata: {})']'
2019-11-08 15:00:57 DEBUG rasa.core.policies.form_policy - There is no active form
2019-11-08 15:00:57 DEBUG rasa.core.policies.ensemble - Predicted next action using policy_1_FallbackPolicy
2019-11-08 15:00:57 DEBUG rasa.core.processor - Predicted next action 'action_listen' with confidence 1.00.
2019-11-08 15:00:57 DEBUG rasa.core.processor - Action 'action_listen' ended with events '[]'
2019-11-08 15:00:57 DEBUG rasa.core.agent - Deleted lock for conversation 'default' (unused)
I was unable to understand that, please try asking in a different way.
Your input -> who are you
2019-11-08 15:01:12 DEBUG rasa.core.agent - Created a new lock for conversation 'default'
2019-11-08 15:01:12 DEBUG rasa.core.tracker_store - Recreating tracker for id 'default'
2019-11-08 15:01:12 DEBUG rasa.core.processor - Received user message 'who are you' with intent '{'name': 'smalltalk.agent.acquaintance', 'confidence': 0.9611070156097412}' and entities '[]'
2019-11-08 15:01:12 DEBUG rasa.core.processor - Logged UserUtterance - tracker now has 6 events
2019-11-08 15:01:12 DEBUG rasa.core.policies.fallback - NLU confidence threshold met, confidence of fallback action set to core threshold (0.3).
2019-11-08 15:01:12 DEBUG rasa.core.policies.form_policy - There is no active form
2019-11-08 15:01:12 DEBUG rasa.core.policies.ensemble - Predicted next action using policy_1_FallbackPolicy
2019-11-08 15:01:12 DEBUG rasa.core.processor - Predicted next action 'utter_default' with confidence 0.30.
2019-11-08 15:01:12 DEBUG rasa.core.processor - Action 'utter_default' ended with events '['BotUttered(text: I am having trouble understanding that, could you please clarify?, data: {"elements": null, "quick_replies": null, "buttons": null, "attachment": null, "image": null, "custom": null}, metadata: {})']'
2019-11-08 15:01:12 DEBUG rasa.core.policies.form_policy - There is no active form
2019-11-08 15:01:12 DEBUG rasa.core.policies.ensemble - Predicted next action using policy_1_FallbackPolicy
2019-11-08 15:01:12 DEBUG rasa.core.processor - Predicted next action 'action_listen' with confidence 1.00.
2019-11-08 15:01:12 DEBUG rasa.core.processor - Action 'action_listen' ended with events '[]'
2019-11-08 15:01:12 DEBUG rasa.core.agent - Deleted lock for conversation 'default' (unused)
I am having trouble understanding that, could you please clarify?
Your input -> what is life insurance
2019-11-08 15:01:56 DEBUG rasa.core.agent - Created a new lock for conversation 'default'
2019-11-08 15:01:56 DEBUG rasa.core.tracker_store - Recreating tracker for id 'default'
2019-11-08 15:01:56 DEBUG rasa.core.processor - Received user message 'what is life insurance' with intent '{'name': 'life_insurance', 'confidence': 0.9586918354034424}' and entities '[]'
2019-11-08 15:01:56 DEBUG rasa.core.processor - Logged UserUtterance - tracker now has 10 events
2019-11-08 15:01:56 DEBUG rasa.core.policies.fallback - NLU confidence threshold met, confidence of fallback action set to core threshold (0.3).
2019-11-08 15:01:56 DEBUG rasa.core.policies.form_policy - There is no active form
2019-11-08 15:01:56 DEBUG rasa.core.policies.ensemble - Predicted next action using policy_1_FallbackPolicy
2019-11-08 15:01:56 DEBUG rasa.core.processor - Predicted next action 'utter_default' with confidence 0.30.
2019-11-08 15:01:56 DEBUG rasa.core.processor - Action 'utter_default' ended with events '['BotUttered(text: I am having trouble understanding that, could you please clarify?, data: {"elements": null, "quick_replies": null, "buttons": null, "attachment": null, "image": null, "custom": null}, metadata: {})']'
2019-11-08 15:01:56 DEBUG rasa.core.policies.form_policy - There is no active form
2019-11-08 15:01:56 DEBUG rasa.core.policies.ensemble - Predicted next action using policy_1_FallbackPolicy
2019-11-08 15:01:56 DEBUG rasa.core.processor - Predicted next action 'action_listen' with confidence 1.00.
2019-11-08 15:01:56 DEBUG rasa.core.processor - Action 'action_listen' ended with events '[]'
2019-11-08 15:01:56 DEBUG rasa.core.agent - Deleted lock for conversation 'default' (unused)
I am having trouble understanding that, could you please clarify?
Hi there, can I ask why you’ve commented out the MemoizationPolicy in your config? It’s a pretty central policy for action prediction. Ideally most of the happy paths would be predicted based on following your stories, while the KerasPolicy is there to jump in when you hit a path you’ve never seen before.
Thanks for the response @erohmensing , After adding the MemoizationPolicy, for the first question only it giving correct response. For remaining questions It is giving default fall back action.
Sorry, I’m back I would first test your stories as you already have to make sure that there are no conflicts in your stories. What do your stories look like? Do you have examples of where the user is asking more than one small talk intent? Perhaps if your stories are all very small and modular, the AugmentedMemoizationPolicy may be the better policy instead of the MemoizationPolicy.