question

MarciStrong-5230 avatar image
0 Votes"
MarciStrong-5230 asked romungi-MSFT answered

How does QnA Maker get smarter based on user behavior?

On https://azure.microsoft.com/en-us/services/cognitive-services/qna-maker/, it says "Your knowledge base gets smarter, too, as it continually learns from user behavior." How does it do this if the bot does not ask users to score responses?
Thank you!

azure-qna-maker
5 |1600 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

1 Answer

romungi-MSFT avatar image
0 Votes"
romungi-MSFT answered

@MarciStrong-5230 QnAMaker uses active learning to improve suggestions and get smarter. This is possible by with the way feedback is handled by its ranker i.e implicit and explicit feedback.

  • Implicit feedback – The ranker understands when a user question has multiple answers with scores that are very close and considers this as feedback. You don't need to do anything for this to happen.


  • Explicit feedback – When multiple answers with little variation in scores are returned from the knowledge base, the client application asks the user which question is the correct question. The user's explicit feedback is sent to QnA Maker with the Train API.

Both methods provide the ranker with similar queries that are clustered.

When this feedback is processed with active learning the model can be trained again and published. I hope this helps!!





5 |1600 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.