search

wit-ai / wit

Natural Language Interface for apps and devices
https://wit.ai/
942 stars 91 forks source link

Intent conflict issue with wit.ai NLP #2728

Open anjaly opened 5 months ago

anjaly commented 5 months ago

Hi, We are a health related application and so we would like to train our bot to answer with some pretrained health related questions. As this is a health-related application, we expect the bot to respond accuratly.

For this we have different questions related to same topic. For example we have some questions related to 'Allergy' and some other related to 'Liver'

Allergy
---------
1. Who should take the Food Allergy Test? 
2. Who should not take a Food Allergy Test? 
3. What is the difference between food allergy and food sensitivity testing? 

Liver
------
1. What are the warning signs of liver disease? 
    2. More common reasons for elevated liver enzymes? 
   3. What can I learn by screening for liver health?

With the node-wit npm, our application UI automatically trains the intent when we input a question. The issue we are facing is that if we train 'Question 1', I'm not able to train the 'Question 2' (returning conflict with 'Question 1') even if both these questions have different meaning with same keywords.

I would appreciate it if you could suggest a solution that allows us to train each question separately. Thanks & Regards

anjaly commented 5 months ago

Hi, We are a health related application and so we would like to train our bot to answer with some pretrained health related questions. As this is a health-related application, we expect the bot to respond accuratly.

For this we have different questions related to same topic. For example we have some questions related to 'Allergy' and some other related to 'Liver'

Allergy
---------
1. Who should take the Food Allergy Test? 
2. Who should not take a Food Allergy Test? 
3. What is the difference between food allergy and food sensitivity testing? 

Liver
------
1. What are the warning signs of liver disease? 
    2. More common reasons for elevated liver enzymes? 
   3. What can I learn by screening for liver health?

With the node-wit npm, our application UI automatically trains the intent when we input a question. The issue we are facing is that if we train 'Question 1', I'm not able to train the 'Question 2' (returning conflict with 'Question 1') even if both these questions have different meaning with same keywords.

I would appreciate it if you could suggest a solution that allows us to train each question separately. Thanks & Regards

Can anyone please suggest a solution for how to tain these questions in wit.ai NLP.

anjaly commented 5 months ago

Author

Hi, We are a health related application and so we would like to train our bot to answer with some pretrained health related questions. As this is a health-related application, we expect the bot to respond accuratly.

For this we have different questions related to same topic. For example we have some questions related to 'Allergy' and some other related to 'Liver'

Allergy
---------
1. Who should take the Food Allergy Test? 
2. Who should not take a Food Allergy Test? 
3. What is the difference between food allergy and food sensitivity testing? 

Liver
------
1. What are the warning signs of liver disease? 
    2. More common reasons for elevated liver enzymes? 
   3. What can I learn by screening for liver health?

With the node-wit npm, our application UI automatically trains the intent when we input a question. The issue we are facing is that if we train 'Question 1', I'm not able to train the 'Question 2' (returning conflict with 'Question 1') even if both these questions have different meaning with same keywords.

I would appreciate it if you could suggest a solution that allows us to train each question separately. Thanks & Regards

Can anyone please suggest a solution for how to tain these questions in wit.ai NLP.

@patapizza can you please suggest a solution for this issue.